Free Papers

Despite progress achieved in the medical treatment of patients with acute pancreatic necrosis (APN) and the role of surgery is no longer in dispute, controversy still exists regarding the ideal time of surgery. We have reviewed our experience at SPUMS between 1980 and 1991 with 68 patients undergoing surgery for APN. Patients with pancreatic abscess were not included. The most common etiologies were biliary(n--25), alcohol (n= 16) and postoperative pancreatitis (n =9). Proportion of males to females was 3.8:1 and the average age was 46 years. The APACHE II score was calculated in all patients and it was higher than 9 in 48% (n=33).

Thirty-three patients were operated before 14 days of illness onset and 35 were operated after the second week onset at the appearance of evidence of infected necrosis. The APACHE II scores for both groups were not different. The surgical predure consisted of necrosectmy plus extended drainage. Reoperation for persistent sepsis occurred in 87.8% (n =29) in the early surgery group patients and in 34.3 % (n = 12) in the delayed treated patients. Infection of the necrotic tissue investigated in 61 patients was present in 83.6% (n =51), 72.5 % of which were monomierobial and was related to early laparotomy in 57% (n=29). The overall mortality rate was 26.4%. In patients with severe pancreatitis (APACHE II > 9) undergoing delayed neerosectmy mortality rate was 18.7% compared to 76.5% mortality associated with early laparotomy.
Necrosectomy delayed until the second week is a suitable procedure and may achieve a low mortality rate in patients with severe pancreatitis with pancreatic necrosis. We investigated 36 patients with pancreonecrosis, using mdionuclear techniques, catheterising the celiac trunk and portal vein. The changes we observed do not correlate with the degree of hypovolemia, but strongly correlates with the fall in cardiac output. In patients with group II severity pancreonecrosis (L.F. Hollcnder, 1) we observed a mean decrease in total blood flow to the liver of 20%. In group III patients we observed a mean decrease in total blood flow to the liver of 53% in comparison with normal individuals. After 3 days of treatment there was only a mean decrease of 41% in total blood flow to the liver in group III patients. While in group II patients the results, were close to the control group. After 6-7 days of treatment the results were very close to those in patients with group I disease. From our data the main reason for decrease in total blood flow to the liver were portal hypertension with increase in pressures of 2,39+_.0,42 kPa (in group II patients) and 3,05_.+0,38 kPa (in group HI patients). The signs of portal hypertension as well as thrombosis of splenic vein and migration of the thrombus were observed at angiography. In our o_pinion hypoxemic damage to the liver played a bigger role than enzymatic aggression, in causing hepatocellular failure, according to biochemical parameters we notice in all patients. We used local fluid infusion through celiac trunk and portal vein for treatment and prophylaxis. This method has improved the outcome of hepatocellular damage and reduce the incidence of paralytic ileus. But there was no reduction in suppurative complications and mortality in patients without operative drainage. Athens, Greece The clinical course of necrotizing pancreatitis and cndotoxaemia show similarities, and, as it is known, the gut contains large amounts of Gram negative bacteria that continuously shed cndotoxins. Using the Limulus Lysatc Test in 1976 we found endotoxin in the ascitic fluid in cases of this pathology. On the other hand enteric micro-organisms arc frequently isolated from different tissues in pancreatitis. All these have raised the hypothesis that the intestinal tract may play a major role in the pathogcncsis of ovcr-whclming septic complications which have been recognised as a primary cause of morbidity and mortality in necrotizing pancreatitis. So, on the basis of this in a session of the Hellenic Surgical Society in 19861 stated: I wonder whether a total colectomy has a therapeutic place in necrotizing pancreatitis .
In November 1990 1 performed the first total colectomy in a lady 82 years old with pancreatitis and in a situation of multiple organ failure. She had a very smooth postoperative course and left the hospital in good condition after three weeks. Since then six more patients have been operated on and a ttal colectomy done. Our own policy is: cholystectomy and decompression of a dilated mmon bile duct if stones are present; total olectomy with an ileostomy and a mucus fistula of the sigmoid colon; scholastic removal of the great omentum and also of the mesxlon from its root. We leave the pancreas alone because the problem is the enzymatic secretion from the "sound" pancreatic tissue and not e necrotic pancreas.
In all seven patients the operative evidence of necrotizing pancreatitis was clearcut. They recovered very soon and five left the hospital earlier than one month. One died from pulmonary embolism 17 days after operation. Another developed an iatrogenic enteric fistula at the time of drainage of a subhepatic abscess and, during the conservative management, a metabolic disorder which had been overlooked also developed and the patient died the 50th postoperative day. Independently of these two deaths, both the results and also that of the origin of endotoxin and bacteria in the gut, strongly support an indication to total colectomy in cases of necrotizing paucreatitis until some effective measures against endotoxin are discovered. FP008 HAEMORRHAGIC COMPLICATIONS IN CHRONIC PANCREATITIS F Callejas Nero, J C Pareja, E A Chaim, S C P Lego, V F Pilla, L S Leonardi D.epartment of Gastroenterology Surgery, University of Campinas, Campmas-SP, Brazil We present our experience in the management of a group of 10 patients who developed haemorrhagic complications due to pseudo-aneurysms secondary to alcoholic chronic pancreatitis. The bleexting was located in pancreatic cysts in four patients, in the digestive tract in four patients and in the peritoneal cavity in two patients. The surgery was on an elective basis in six cases and emergency in four cases. Surgical procedures performed were distal pancreatectomy (50%) and ligation of the artery with pseudo-aneurysm associated with longitudinal pancreaticojejunostomy (50%). There was no operative mortality and no haemorrhage recidive in the follow-up at 36 months. In selected cases the combination of ligation and pancreatic drainage could be performed at the same time to manage the pain and the haemorrhagic complication of chronic pancreatitis.  (17),was provided as follows.ln 5 cases the cysts were totally excised, with relative ease, but in another 5 patients atypioa! hepatectomy was inevitable. The residual cyst space in the remaining 7 was managed by removal of all living cyst elements, llgation of the leaking bile radicles,sterilization of the cavity and drainage by tube placement. 13 of the 33 complicated csts had ruptured into the bile ducts necessltatlng,geyond the arorementloned treatment,as for slmpe oss,choleoyste otomy, bi|e duct exploration and T tuoe insertion. In all patients with ihfeoted oysts (20),the cavities were treated as it is mentioned above,but we proceeded with omentoplas.ty and drained both the ,residual space,by lacing a three way Folley catheter at the subhepatio ouoh, y fixin an elastic tube (pen-rose).
From 1985 to 1991 we treated 83 consecutive patients with HH, exclusively by drainage operations. (Isolation of the peritoneal cavity, wide unroofing of the cyst, careful evacuation of the cyst contents, sterilization with hypertonic saline, interlocking haemostatic suturing of the cyst edges and postoperative drainage with closed drainage systems).
There was postoperative mortality of 1% (cerebrovascular accident) and morbidity of 12% (two abscesses, eight biliary fistulas). One abscess was treated by surgical and one by percutaneous drainage.
Three fistulas were treated expectantly (healing required three to 12 months) and five by endoscopic sphincterotomy and nasobiliary tube drainage (healing required 7 to 14 days). No statistical improrement was shown by adding an omentoplasty. There was a recurrence rate of 8% (follow up from one to seven years, achieved in 50 patients). A combined application of sheets of collagen covered with freshly prepared fibrin glue improved local haemostasis to a great extent. Large areas of capillary bleedings can be treated successfully. But due to the relatively complicated preparation required at the operation site, this method has not been used on a large scale. These drawbacks have been overcome with the latest development in this field a sheet of collagen covered with a fixed layer of the solid components of fibrin glue (fibrinogen, thrombin and aprotinin).
In a prospective study, 225 cases of liver resections due to metastases (43 %), primary liver carcinomas (16 %), intraoperational injuries of Glisson's capsule (10 %), liver ruptures, benign tumours, bleeding subsequent to punch biopsies, echinococcus cysts and others were included. The assessment of the haemostyptic properties of TachoComb was "very good" and "good" in 95 %, "satisfactory" in 4 % and unsatisfactory in 1% of the cases.
In several cases, particularly difficult bleeding situations could be controlled, e.g. those with massive coagulation disturbances or hepatic stasis. No complications attributable to TachoComb occurred. So far, more than 280 patients with liver resections have been treated successfully with TachoComb in clincal trials. In a liver transplantation study with 27 patients, haemostasis with TachoComb was assessed as "very good" and "good" in 96 % of the cases. Liver resection are often associated with operative bleeding and blood transfusion. The purpose of this work was to determine which patients were particularly at risk of intraoperative blood transfusion. Between October 1990 and July 1992, 100 patients had liver resection for a liver tumor hepatocellular carcinoma in cirrhosis (37 patients), primary liver cancer in normal liver (14 patients), liver metastases without previous chemotherapy (16 patients) or following chemotherapy (3 patients) and benign liver tumor (30 patients). There were 12 atypical resections, 39 segmentectomies, 25 major hepatectomies, 15 extended hepatectomies and 9 liver resections associated with portal, IVC or biliary reconstruction. Clamping of the portal triad was used in all patients and hepatic vascular exclusion in one patient with IVC replacement. Twenty two (22 %) patients required blood transfusion. It was significantly (p < 0.01) more frequent in patients with cirrhosis (41%) or following chemotherapy (66 %) than in patients with a normal liver (4.5 %). The rate of blood transfusion was almost similar during atypical (17 %), segmental (5 %) or major (16 %) liver resection. It was slightly higher after extended (40 %) and significantly higher during resection with associated procedures (89 % p < O.O1). The mean volume of blood transfusion per transfused patient was around 5 units of blood whichever tumor was resected and whichever resection was performed. These results suggest that blood transfusion is not frequent in patients will a normal liver undergoing liver resection when no associated procedure is performed. Autotransfusion, hemodilution, hepatic vascular exclusion (HVE) are not required in such cases. Blood transfusion is frequent and abundant in cirrhotic patients and patients with chemotherapy undergoing an extended liver resection or a liver resection associated with reconstruction of the portal vein, IVC or bile duct. Only in such patients a combination of HVE, hemodilution and autotransfusion might decrease transfusional needs.  MATERIAL and METHODS In anatomical liver resections the intrahepatic vascular anatomy and tumors have to be related to the external landmarks on the liversurface in order to define the precise extent of the planned resection. The relation between internal anatomy and external landmarks on the liversurface was studied in 10 voluneers using 3-D reconstructions of MRI acquisitions. RESULTS Two out of 10 subjects showed a normal pattern of 3 hepatic veins, middle and left vein sharing a common runk. The remaining 8 subjects had 9 accesory hepatic veins, 4 right, 2 middle and 3 left. Additionally, I subjec had a separate origen of he left hepatic vein and righ hepatic vein. Two out of I0 subjects showed a conventional portal branching pattern with a separate righ anterior and posterior trunk. Two subjects had a separate branch to the right liver which originated from the main portal vein before it's division in right and left portal trunk. subjects showed a common trunk to segment 5 and 6 and 4 ohers showed a trunk to segment 7 and 8. On average the gallbladder was located 6 degrees to he right of the middle hepatic vein (range 1-18). The righ hepatic vein was allways located in the coronal plane or slightly posterior ro i. The transverse fissure could not be assigned in 3 our of 10 patients due to the presence of more than one right portal trunk. CONCLUSION Marked individual variations in hepatic and portal vein anatomy, known from in vivo and in vitro studies, can be shown with 3-D liver imaging thus providing the possibility of an individualized preoperative planning procedure. In 409 patients with insulin-dependent diabetes mellitus, distal splenorenal venous shunt was made to divert pancreatic blood rich with hormone glucagon into the general circulation to meet injected insulin (ins) in the peripheral tissues. After the operation the ratio of exogenous insulin and endogenous glucagon in the tissues (hepatic and pcriphcric) seems to bc more optimal for their interaction. Angiographic control was made in 137 of them at a period of 8 months after surgery and included renal vcnography, selective splenic venography. In 114 patients (83 %) splcnorenal shunt proved to be permeable and in 23 patients (17%) opacification of the splenic vein was not achieved after using all the above methods.
After the shunting operation, the patients with permeable anastomosis showed disappearance of, or diminishing complaints of leg pains, general weakness, clinical manifestation of hypo-or hyperglycaemia. The dose of injected exogenous ins. reduced from 0,95 + 0,05 Un/kg to 0,76+ 0,04 On/kg (P < 0,05) without changing the preparation and schedule of injections. In 16 patients improvement took place without changing the dose of exogenous ins.
The level of glycosilated haemoglobin (HbAc) was 9,3+ 0,3 % against 13,3+ 0,3 prior to surgery (P < 0,05). At the same period of time after the operation patients with non permeable anastomosis had the same complaints and the same dose of ins. as before surgery. The level of HBAc remained high" 12,8+_.0,3%. It was proved that development of splenorenal collaterals in patients with permeable anastomosis reduced the efficiency of the shunting procexiure, therefore we now perform additional ligation of vv. gastrica sin. and gastro-epiploica sin. Hepatic function did not suffer after surgery.
Stabilization of clinical state and metabolism after the operation and dependence of remote results on permeability of applied anastomosis wimcss efficiency of the new approach to treatment of insulin-dependent diabetes mcllitus. Local and remote tissue injury results from the development of acute pancreatitis, in particular, the sequestration of neutrophils (PMNs) occurs at sites of capillary endothelial cell injury in the lung and pancreas. Other data from our laboratory indicate a pathogenic role for these PMNs. It is not known whether this PMN sequestration results from specific humoral factor(s), or from alterations in the target endothelial cells, or both. This study was designed to assess whether circulating plasma in animals with acute pancreatitis contains humoral factor(s) which enhance PMN adhesion. METHODS: Secretagogue-induced acute pancreatitis was generated in 200 gram Sprage Dawley rats by continuous intravenous infusion of supraphysiologic doses (5tg/kg/min) of caerulein, a CCK analog. Blood was obtained after 3 hours by venipuncture from the inferior vena cava and plasma PMN aggregation potential was quantitatively determined electronically with a Coulter TM Counter by a method previously developed and described. Plasma from rats with caerulin-induced acute pancreatitis has significantly greater leukoaggregafion potential than plasma derived from saline infused control animals. These data offer compelling evidence for the presence of humoral factor(s) which enhance PMN adhesion following acute pancreatitis.
The nature of these factors is the subject of ongoing investigation. In a controlled prospective randomized trial serumendotoxin levels were measured after induction of acute experimental pancreatitis and synchronous colonic irrigation or creation of a cecostomy. Thus, after inducing acute pancreatitis in male wistar rats by injecting a 2% taurocholate-solution into a temporarily closed duodenal loop according to a technique first described by Orda and coworkers (Arch. Surg. 1980) 40 animals underwent cecostomy (group B), in 36 the colon was irrigated with normal saline (group C) and 36 served as controls (group A).
Altogether, twelve animals succumbed intra-or immediately postoperatively. As a consequence 31 remained in group A, 36 in group B, and 33 in group C. Before the end of scheduled follow-up (15 days) the mortality rate in the control group was 22.6%, 16.7% in group B and 9.1% in group C, with the difference between group A and C being statistically significant at p < 0.05. This difference was parallelled by a significant difference of serum endotoxin levels between the groups (219 ng/l in group A, 79.2 ng/l in group B, and 71.7 ng/l in group C). Also the number of serum endotoxin positive animals was significantly higher in group A than in group B and C (7 in group A vs. 2 in group B and C). Our results suggest that endotoxin might play an important role in the pathophysiology of acute pancreatitis and that endotoxin absorption might be reduced by cecostomy or colonic irrigation. The essential fatty acid, linoleic acid, constitutes about 60-70% of the clinically available lipid emulsions, which are now routinely provided to patients receiving parenteral nutrition to avoid the development of essential fatty deficiency. However, it has been shown in experimental models, that the infusion of polyunsaturated fatty acids (PUFA) into rats and isolated perifused islets, will stimulate insulin secretion, but will render the beta cells unresponsive to glucose. The mechanism of both the stimulatory and desensitization effects of PUFA was shown to be linked to fatty acid oxidation. The AIM of the present study was to explore the possibility of restoring the glucose effect by the provision of L-glutamine, a major fuel soume and a precursor for the biosynthesis of the antioxidant, glutathione (GSH) to the PUFA perifusate of isolated islets. METHOD: In each experiment, a batch of six islets microdissected from three female CD-1 mice were preperifused for 1 hour at 37C at the rate of l ml/min with a Krebs-Ringer bicarbonate buffer, containing 5.SmM glucose (basal), 2% bovine albumin, 100 KIU/ml trasylol and maintained at pH 7.4 by continuous gassing with 95%/5% Oz/CO z. Basal effluent samples were then taken before the glucose concentration was raised to 27.7mM for 20 minutes; immediately followed by 20 mins basal glucose 'washout' in the absence or presence of 20raM glutamine (GLN) or 3raM glutathione (GSH). Linoleate (10raM) was then added to the basal perifusate without or with GLN or GSH. Effluent perifusate samples were collected on ice at 2 rain-intervals and stored frozen until radioimmunoassay for insulin. RESULTS: The stimulatory effect of 27.7mM glucose on insulin secretion assessed as the incremental area under the curve were 1552.8 _+ 276.3 and 220.4 +_ 163.9 pg/20 mins (p<0.001, n=6) respectively, before and after linoleic acid treatment alone. In experiments, in which islets were treated with the fatty acid in the presence of L-glutamine, there was no difference in the insulin response to 27.7mM glucose before and after linoleate treatment (2051.8 +_. 420.5 vs 2159.2 pg/20 rains respectively, n=6). When GSH was substituted for L-glutamine, qualitatively similar results were observed. CONCLUSION" The presence of glutamine or GSH completely blocked the linoleate-induced desensitization of beta cell secretory response to glucose, suggesting similar a mechanism between their actions. These data support an efficacious role for L-glutamine during hyperalimentation. At the present time surgical resection is the sole chance of cure in pancreatic cancer but only for 5-10% of the patients. Efficiency of surgical resection could be perhaps enhanced by an adjunctive intraoperative photodynamic treatment. The aim of this work was to study the survival of a model of rat pancreatic cancer treated by Pheophorbide A (PPA) photodynamic therapy (PDT).
The tumoral model was an acinar pancreatic cancer induced by azasedne and transplanted in the pancreatic tail of Lewis rat. The survival time of this model do not exceed 33 days. The.photosensitizer used (PPA) was a chlorophyll derivative at a dosage of 9 mg. kg-1.intravenously administered 24 hours before illumination and achieving a ratio of 12:1 between tumor vs surrounding normal pancreas. A DCM laser pumped by a copper vapor laser was used to deliver a fluence of 100 J.cm-2 at 660 nm. The spot size of laser light was 3 cm-2. The tails of the pancreas (bearing a tumor of 1 cm in diameter) of 18 rats (9 with PPA and 9 without) were exposed under a midline laparotomy. The surface temperature of the tumor was monitored throughout the illumination. The maximum tumor tenperature rise was less than 2 C. This elevation was not considered in favour of thermal effects. Rats follow up was established according with Kaplan Meier method. Autopsies were performed on dead rats. Rats surviving more than 4 months were sacrificed for histological examinations.
Control rats (9 illuminations without PPA, 9 PPA without illumination and 9 without any kind of treatment) did not survive more than 33 days with evidence of metastatic spread and carcinomatosis. At this time, the rate of survival in the PDT group was 89% (log rank test, p < 0,001). Six animals (66%) were still alive 4 months after PDT. Histological examination of the survivals showed evidence of pancreatic tail necrosis without metastatic spread.
Further experiments were conducted histologically in order to determine the threshold of PDT damage. 33 and 75 J. cm-2 induced a surrounding edema of pancreatic acini but failed to destroy tumor cells. Tumor destruction was achieved only at a fluence of 100 J. cm-2. It should be observed that normal pancreas was immune to PDT damage even at a fluence of 100 J. cm-2.
The efficiency of photodynamic therapy using pheophorbide A was assessed not only by histological selective necrosis of pancreatic tumor but also by 2/3 survival of PDT rats. This work suggests that intraoperative adjunctive PDT may improve therapeutic results of pancreatic carcinoma surgery. The use of optical spectroscopy is a potential new approach for the diagnosis of pancreatic malignant mmours. We will report the use of laser induced fluorescence of a new photosensitizer (PPA) which presents tumour-localising properties.
The tumoral model used in this experiment was an acinar pancreatic cancer induced by azaserine and transplanted in the pancreatic tail of Lewis rat. The PPA is a chlorophyll derivative used in our experiment at a dosage of 9 mg. kg-1 IV 24 hours before illumination. Fluorescence emission spectra of 5 tumours and their surrounding pancreas obtained under 400 nm excitation showed a broad double peak with a maximum at 678 nm. Autofluorescence from tissue was also observed at 463 nm. A reduction of PPA fluorescence intensity was found at the surface of the tumours because of their heavy blood staining. Consequently, PPA fluorescence signal alone was unable to provide a photodynamic ,image of pancreatic carcinoma.
In order to avoid problems related to variations in distance, surface topography, drifts in laser intensity, detection efficiency and overall cancer blood staining, we have developed a concept based on a dimensionless function. This function was established by dividing the intensity fluorescence signal of PPA by autofluorescence. We called it Rt for the mmour and Rp for the surrounding pancreas. The fluorescence contrast C = Rt I Rp should be greater than one to allow obtaining of PDI of a tumour.
To determine the best excitation we have tested the previously established excitation wavelengths of PPA (355, 400, 510, 530 and 610 nm) on 5 tumours and surrounding pancreas. Obviously, 355 nm gave the best fluorescence contrast (C = 1_) and was used to perform.our preliminary imaging of an intrapancreatic tumour and its intraperitoneal metastasis.
We used a Nd" YAG laser at 355 nm as excitation source. Fluorescence was recorded by a CCD camera through 2 interferential filters: in the red for PPA signal and in the green for autofluorescence. A dimensionless contrast function has been calculated for each spatial location using the values in corresponding pixels in the two images. A resulting artificial image was formed with false colours coding.
High contrast images have been obtained from the tumour and peritoneal metastasis.
PDI of pancreatic carcinoma using PPA as a dye seems to be effective. Such an imaging process may perhaps allow in the future to detect and cure the metastatic spread in the time course of debulking surgery of a pancreatic carcinoma. We have shown previously that under nutrition reduces the mortality of acute experimental pancreatitis by decreasing the pancreatic enzyme content.
Cerulein in physiologic doses reduces the enzyme content of the pancreas without any harmful effect t the pancreas.
The aim of this study was to assess the effect of acute reduction of pancreatic enzyme content by using inframaximal doses of cerulein in the outcome of acute pancreatitis (AP).
Method: Two groups of rats were studied: Group Io con.trol animals fasting 12 hours and Group IIanimals receiving cerulein in inframamal dses (0,2g/kg/h) with free access to water and food. Enzymatic content of the pancreas was studied in both groups. Acute pancreatitis was induced by retrograde injections of 0,5 ml of 2,5 % Na-tauroclate into the pancreatic duct.
Conclusions: These results indicate that decreasing enzymatic content of the pancreas reduces the severity and gives a protective effect in acute pancreatitis in rats. Compcions occurred in 8 patients (28.5%) from the T-T group and in 4 patients (13.3%) from the NT-T group. In the T-T group 7 had early (<3 months) complications including: accidental dlslodgement (2), bile leakage (3), cholangltls from bile stasis (I) and T-tube blockage (I). There was a late complication (>3 months): localized peritonitis after T-tube removal. In the NT-T group 4 patients experienced bile duct stricture: one early and 3 in the late postoperative period. All patients were treated with endoscopic baloon dilatation but 2 required surgical reconstruction and one was retransplanted because of hepatic artery thrombosis. Fifty two patients survive a mean follow up of 5.7 months. Six patients died (10%) from causes unrelated to billary complications.
In biliary anastomosis during OLT there was significant increase in morbidity with the use of a T-tube, when compared to those without a T-tube stent (28.5% vs 13.3%). Bile leaks occurred in 38 cases due to anastomotic leaks ( 17 ), 9 of whom subsequently developed strictures, asymptomatic leak diagnosed on protocol T tube cholangiography (9), leak following hepatic artery thrombosis (4) leak following T-tube removal ( 5 ), cystic duct leak ( 2 ), segmental hepatic duct leak (i). Biliary obstruction occurred in 48 cases due to anastomotic stricture (7), intrahepatic stricture ( 9), papillary stenosis/ recipient duct obstruction (8) Improval in survival following liver transplantation has brought to light late complications, particularly those concerning the biliary tract. The aim of this study is to determine the risk factors of these late biliary complications.
Between April 1978 and June 1992, 135 orthotopic liver transplantations were performed. During the harvesting procedure (rapid technique as of 1978) the bile ducts were rinsed by injection of saline via the gallbladder. The anastomosis was choledoco-choledochal, choledocojejunal and cholecysto-jejunal in 102, 27 and 3 cases respectively. Three patients died intraoperatively, prior to biliary anastomosis. Among ll0 functional grafts at 1 month, 12 late biliary complications were observed (10.9%). Six of these late biliary complications consisted of an association of strictures and dilatations of the intrahepatic bile ducts. Among the group (n=12) with late biliary complications, 4 showed stenoses of the hepatic artery or one of its branches, whereas only one of the patients in the group (n=98) without late biliary complications showed arterial lesions (p<0.001). Excluding patients with hepatic artery stenosis or thrombosis, mean cold ischemia time of liver grafts that developed late biliary complications was 859+259 minutes versus 583+283 minutes for those having not developed biliary lesions (p<0.01). Late biliary complications were not correlated with the type of biliary anastomosis or the presence of a positive cross-match. Seven reoperations and 1 retransplantation was performed for late biliary complications.
These results suggest that prolonged cold ischemia times and reduced arterial inflow increase the incidence of late biliary complications. Combined heart-liver transplantation had been performed sporadically with mixed success. Since the first case was performed in 1984, the procedure has met with severe intraoperative difficulties such as haemorrhage and cardiac dysfunction. The main reason for these had been the simultaneous implantation of the two organs under cardio-pulmonary bypass. In order to circumvent this a staged procedure was described where the liver was grafted 2 weeks after the heart. The obvious disadvantages are that the organs were from different donors and two major surgical procedures were required.
We described a modification of the technique whereby the organs were grafted sequentially in a single operation. This was performed in two patients with great success. The first patient was a 33 year old woman with familial hypercholesterolaemia and severe coronary artery disease, the other was a 60 year old man with familial amyloidotic polyneuropathy in moderate heart failure. The latter was the first time such an opeation was performed for this condition. We encountered no intraoperative problem and except for a mild rejection in the first patient their postoperative recovery was uneventful. Bleeding from oesophageal ulcers, oesophagitis or from the varices themselves after injection sclerotherapy is occasionally massive and difficult to control. Since our current experience with octreotide suggest that it is a safe and effective treatment for the control of the acute variceal bleed, we have examined its efficacy in these post-sclerotherapy problems.
Haemorrhage was successfully controlled by an infusion of octreotide in all patients with oesophagitis, in 30 of 31 patients with oesophageal ulceration and in 38 of 42 patients with bleeding from varices. In the 1 patient with persistent haemorrhage from ulcers and in 2 of the 4 with continued bleeding from varices, haemostasis was achieved by hourly boluses of 50 /g octreotide for 24h. No complications were associated with octreotide administration.
The results of this study clearly indicate that octreotide is a safe and effective treatment for the control of the severe haemorrhage after technically successful injection sclerotherapy.  IS MORE COST  EFFECTIVE  THAN  OESOPHAGOGASTRIC  DEVASCULARIZATION AND TRANSECTION IN THE LONG-TERM MANAGEMENT OF BLEEDING OESOPHAGEAL  VARICES: INTERIM ANALYSIS OF A PROSPECTIVE  RANDOMIZED CONTROLLED TRIAL. JEJ Krige, PA Goldberg, PC Bornman, J Terblanche. Department of Surgery and MRC Liver Research Centre, University of Cape Town and Groote Schuur Hospital, Observatory, Cape Town, South Africa.
Fifty-two patients (35 male, mean age 45.8 years, range 18 65 years) with variceal bleeding were randomised after emergency endoscopic sclerotherapy to continued endoscopic sclerotherapy (ES) using 5% ethanolamine until variceal obliteration followed by regular check endoscopy (n--27) or to oesophagogastric devascularization with transection (OGDT;n =25). Childs C score > 11, those over 65 yrs and high risk operative patients were excluded. Thirty-seven pts had alcoholic cirrhosis; 7 were Childs A, 26 Childs B and 19 Childs C. Mean follow-up was 19 months (range 6-58 months).
All data were analyzed on an intention to treat basis. Mortality during the first month after randomization was higher in the surgical group (2/25 v 0/27) but late deaths in the OGDT pts were fewer than among the ES group (5 vs 9). Varices were eradicated in 24 of 27 patients in the ES group after a mean of 5 injections (range 2-10). Three patients in the ES group died before eradication at a mean of 96 days. One patient in each group required dilatation for an oesophageal stricture. No patient died from rebleeding.
During follow-up there were no significant differences between the ES and OGDT groups with regard to number of patients rebleeding from varices (6 vs 7), number of bleeding episodes (8 vs 6) number of units of blood transfused per patient (0.7 vs 0.4), total number of hospital admissions (79 vs 70), mean number of days in hospital per pt (52 vs 45) or mean days per admission (15 vs 16). The mean cost per patient was however significantly more in the group undergoing operation (OGDT--$12,475; ES = $7,542). We conclude that while ES and OGDT are equally effective in eradicating varices and preventing rebleeding, sclerotherapy is significantly more cost effective than operation. The purpose of this work was to compare the results of Sugiura procedure (SP) and portacaval shunt (PCS)in the elective treatment of cirrhotic patients with previous variceal bleeding. Fifty-four patients were included in the study between January 1986 and April 1989. Twentyseven patients were randomized into the PCS group and underwent a direct side-to-side portacaval shunt (22 patients), an interposed portacaval shunt (1 patient), or a mesocaval shunt (4 patients). Twentyseven patients were randomized into the SP group and had a Sugiura procedure (26 patients) or a mesocaval shunt (1 patient). The intention to treat principle was applied to this study. The two groups were comparable according to etiology of cirrhosis, liver function tests, number and severity of bleeding episodes and size of varices. Total and variceal recurrent bleeding episodes were more frequent after SP (33 % and 22 %) than after PCS (11% and 4 %). The difference however was not significant. In each group, one patient died from recurrent variceal bleeding. Encephalopathy occurred in 8 patients after SP and in 15 patients after PCS. The rate of chronic encephalopathy was significantly higher after PCS (40 %) than after SP (7 %, p < 0.05). One-, 2-and 3year survival (Kaplan-Meier) were respectively 93 %, 81% and 67 % after SP and 78 %, 66 % and 39 % after PCS. The difference between survival curves (Log-Rank) was significant (p < 0.01). These results suggest that, although it is slightly less efficient in preventing recurrent bleeding, the Sugiura procedure is better tolerated and associated with longer survival than portacaval shunt. The Sugiura procedure is an operation of choice when surgical prevention of recurrent variceal bleeding is contemplated. In end-stage cirrhosis complicated by variceal hemorrhage, treatments such as portosystemic shunts which aim to reduce portal pressure decrease sinusoidal perfusion at the same time, with the risk of impairment of liver function. A new concept was put forward that encouraging portal flow to pass through the cirrhotic liver by mechanical action could result in a decrease of distal (splanchnic) portal pressure on one side, and improvement of liver function on the other side. The aim of this work was to evaluate the hemodynamic and functional effects of a short term pump driven increase of portal blood flow through the liver of 13 patients with end-stage cirrhosis before the anhepatic phase of liver transplantation. In 10 of them, portal flow was increased peroperatively during 30 minutes, using a pump-driven venovenous bypass in a portoportal closed circulation. Basal portal flow (800 + 270 ml.min-1) was increased two-fold (n--lO) or four-fold (n--9). When flow was doubled, splanchnic portal pressure decreased 18% (from 31.8 + 5.7 to 26.0 + 5.8 mmHg, p<O.001); when flow was increased four fold, splanchnic portal pressure decreased 39.2 + 15.4% (from 32.8 +/-5.0 to 19.9 + 6.0 mmHg, p<0.001).
Comparison of indocyanine green clearance between basal and doubled portal flow demonstrated an increase of 32.1 +/-26.9% (n--S; p-0.OS3). Histological analysis demonstrated sinusoidal dilatation in 3 out of 10 livers. According to these results, and with previous studies using isolated perfused cirrhotic rat or human livers, we suggest that portal pumping should be explored as a new perspective of treatment for some cirrhotic patients, sclerotherapy-reststant, with variceal hemorrhage and liver failure. Marginal resection of 150-200g of liver tissue and intra-operative application of He-Ne laser radiation (670rim) for 10 min; ligation of splenic artery and omentohepatopexy; resection of 300cm of peritoneum on both sides of the abdominal wall peritoneal-muscular shunt.
The main purpose of the first procedure was stimulation of liver regeneration; the second decreasing portal hypertension. The last part of the operation was made for the reduction of ascitic liquid in the peritone cavity. The mortality rate at the first month after operation was 8.8%. Follow-up results were investigated for 1-5 years after operation (31 patients, group A) in comparison with non-operated patients with Child C liver cirrhosis (35 patients, group B). 41 gallbladder strips responded to EFS; 44 did not. In responders the median inflammation score was 7 (range 3-12) and in nonresponders it was II (5-16; p<0.001). All gallbladders responded to CCK-OP. There was a significant (p=0.0032) inverse relationship between inflammation score and increasing tension.

Number of survivors present in the
The EFS stimulation characteristics selectively stimulate nerves. These results show that neurally mediated contraction in the gallbladder is likely to be lost in the presence of extensive inflammation. Contractility in response to CCK was reduced by increasing inflammation. These findings may be related to loss of nerve tissue or impaired muscle cell function in inflammation. Endoscopic sphincterotomy leaving the gallbladder 'in situ' (EE-GiS) has emerged as an useful alternative to surgery, but little interest has been paid to gallbladder function after ES, and several studies report contradictories results. AIM: To study the modifications in gallbladder function after sphincterotomy in an experimental model of lithogenesis in the rabbit. MATERIAL AND METHODS: 38 male New Zeland rabbits were used. Group I, Control. Group II, Surgical sphincterotomy (SE). Group III, Gallbladder lithiasis induced by a lithogenic diet (colestanol 5%) during 2 weeks (Group IIIa) and 6 weeks (Group IIIb). Group IV. SE after induction of gallstones. Group V. SE previous to the induction of gallstones. A biliary scintigraphy with HIDA, weight of the gallstones, hepatic and gallbladder bile composition and hepatic blood function test (GGT, Alkaline Phosphatase (APh) an Bi) were measured. RESULTS: HIDA observed the gallbladder filling in all cases with a increased excretion fraction in the SE animals. In all the groups, gallstones were induced, but SE precluded the formation of gallstones in Group V. Dry weight of the stones was lower in the SE animals than in controls. Hepatic function blood test were in normal range in all cases. (Table I) Bedford Park, South Australia 5042. Balloon distension of the duodenum can provoke a change in sphincter of Oddi activity. The mechanism(s) which mediates such a response has not been defined. The aims of this study were to determine if: (i) electrical field stimulation of the duodenum influences sphincter of Oddi activity (ii) this response is neurally mediated (iii) these pathways are intramural and (iv) nicotinic and/or muscafinie receptors are involved. Twenty eight anaesthetized Australian Brush-tailed possums (Tfichosurus vulpecula) were used. Electrical field stimulation [70V, 0.5ms, 5-60Hz, 10-20s] on the anterior serosal surface of the duodenum at 2, 3, 4, and 6 cm proximal or distal to the sphincter of Oddi, was used to stimulate neural pathways. The sphincter of Oddi phasic contractions were recorded by manometry. Tetrodotoxin (91.tg/kg) was administered by close intraarterial injection to achieve neural blockade. Hexamethonium (30 mg/kg) and atropine (30 I.tg/kg) were administered i v. sphincter of Oddi activity was quantified by measuring the area under phasic contractions (ram 2 per 30 see). The response was expressed as a % of the pre-stimulus activity.
All possums displayed spontaneous sphincter of Oddi phasic contractions. Electrical field stimulation of the proximal and distal duodenum produced excitatory sphincter of Oddi responses in all animals. These responses were frequency-dependant and maximal at 30Hz. The responses were produced when the duodenum proximal to the sphincter of Oddi (up to 4-5 era, not at or beyond the pylorus n=4) or distal to the sphincter of Oddi (up to 4 era; n=4), was stimulated. The responses were abolished by either pretreatment with tetrodotoxin (n=4), or transsection of the duodenum between the site of stimulation and the sphincter of Oddi (n=3). Hexamethonium did not significantly alter the sphincter of Oddi response to proximal (n=6) or distal (n=8) duodenal stimulation. Atropine reduced sphincter of Oddi response to proximal duodenal stimulation by 58.5 + 8.5 % (n=6) and to distal duodenal stimulation by 34.0 d: 9.1% (n=8) (both P<0.03, Wilcoxon Test).
In conclusion, we have demonstrated the existence of intramural neural pathways between the sphincter of Oddi and the segment of duodenum 4cm proximal and 4era distal to the sphincter of Oddi. These postganglionic pathways involve muscarinic receptors. The increased incidence of gallstones after truncal vagotomy has been attributed to bile stasls. CCK is known to be the principal hormonal mediator of gallbladder contraction. Our aim was to study the effect of truncal vagotomy on gallbladder contraction and CCK levels in response to a meal.
We studied 7 patients after truncal vagotomy and pyloroplasty and 13 normal subjects. After an overnight fast, gallbladder volume was measured in the fasting state and at 15,30, 45, 60 and 90 minutes after a solid fatty meal using the Ellipsoid formula (Dodds). CCK was measured by a specific radlolmmuno assay at the same times. Gallbladder contraction was triphasic in both groups with a phase of relaxation separating two contraction phases.
Plasma CCK was significanty elevated after truncal vagotomy and may contribute to post vagotomy symptoms. There is decreased contractility of the gallblader with a lower ejection fraction and a higher residual volume after truncal vagotomy with consequent bile stasis. These findings suggest that reduced gallbladder emptying after truncal vogotomy diminishes feedback inhibition of CCK release. We had reported at the 4th World Congress of HPB Surgery that free radical (FR) appears in gallstone in vivo, and that the presence of FR was essential in pigment gallstone (PS) formation. The AIM of this study is to explore the effect of FR on the precipitation of calcium bilirubinate (CAB) and glycoprotein (GP), the main ingredients of PS, from human bile in vitro.

60
FR intensity of bilirubin was enhanced by Co radiation (BrE) or reduced by ascorbic acid and mannitol treatment (BrR). When this was confirmed by electron paramagnetic resonance spectra, 0.5 ml of BrE or BrR solution (Smg/ml, pH=8.4) was separately added into two 5 ml aliquots from each of 14 T-tube bile samples. After 36hour incubation, the weight of dried precipitate, amount of GP and CaB in it, contents of GP, total bilirubin (TBr) and total calcium (TCa) in supernatant were measured and compared between paired aliquots (cf. Long term duodenogastric reflux has been implicated in the genesis of gastric stump carcinoma after previous partial gastrectomy for a benign disease. The aim of this experimental study is to assess the effect of a permanent biliary (BR), pancreatic (PR), and bilio-pancreatic reflux (BPR) on gastric mucosa After 2 months exposure, gastric mucosa did not display any macroscopical changes wathever the kind of reflux. On histological examination no intestinal metaplasia, hyperplasia or dysplasia were found. The only lesion found was the presence of numerous intra mucosal cysts nearly the anastomosis. The mean number of intra mucosal cysts was 119 for BR (n=5), 106 for PR (n=6), 60 for BPR (n=5) and 13 for control group (gastric suture n=5) Long term exposure (one year) was investigated on BPR (n=9). Macroscopic examination showed a protruded lesion on the anastomosis of 8 rats. On histologic examination these tumours were in 6 cases an adenocarcinoma and in 2 cases a benign submucosal adenocystic proliferation. The six adenocarcinoma occured in five cases on an adeno-cystic proliferation, and in one case on an inflammatory granuloma.
Our results showed a clear corelation between long term exposure of BPR and gastric carcinoma in rats. Histologic investigations suggest that carcinoma arise through adenocystic proliferation, condtition similar to human gastritis cystic polyposa considered as a premalignant disease. This experimental model could be usefull for other investigastions. This study was performed at the University of Erciyes in Experimental Research Laboratory between November 1991 and August 1992; to investigate the effect of soliidal agents, especially silver nitrate 0.5%, on liver and biliary ducts and to find out whether these agents caused secondary slerosing cholangitis or not. Three groups rabbits were taken in this study. Each group was included 15 rabbits. After laparotomy; The first group of rabbits was given sodium chloride 0.9% into biliary tract. The second group of rabbits was given silver nitrate 0.5%, and the third group of rabbits was given formaline 5%. Blood samples for SGOT, SGPT and ALP analises and liver wedge biopsies were taken. Five rabbits from each group were sacrified at the end of the first, fourth and eighth weeks. Liver, gall bladder and choledo biopsies for histopathologic study were done. There was no histopathologi and biochemical change in the first group. The eighth week ALP values were significantly different from the beginning values in the formaline group (p < 0.05). The first and eighth week SGOT values and the eighth week ALP values were significantly different from the beginning values in the silver nitrate group (p < 0.05). In the both of formaline and silver nitrate groups, some of macroscopic findings and more of microscopic findings of secondary sclerosing cholangitis were found.
As a result as formaline was found responsible for sclerosing cholangitis earlier, silver nitrate 0.5% was also found responsible for slerosing cholangitis. Therefore, In the hydatid disease, any sclerosing scolicidal solution should not be given into the cyst cavity, till the solution must be proved not to be responsible for sclerosing cholangitis. reported after surgical treatment of hydatid disease of the liver, and has been hypothetically attributed to the caustic effect of the scolicidal solution diffusing into the biliary tree trough a cystic-biliary fistula. The aim of this study was to assess the effects of scolisidal solutions on biliary epithelium. In 146 rats we have injected directly into the biliary tract 0.15 ml of isotonic saline (control group), 20% hypertonic saline, 0.5%, and 2% formalin.
Histologic investigations performed 3 months after injection showed no change in biliary structure in the control group (n=11). In hypertonic saline (n=14), and 0,5% formalin (n=12) groups, we observed very mild change of biliary epithelium, and no fibrosis. After 2 % formalin injection (n=16), there was periductal sclerosis in 11 rats and 4 of them developed pseudocirrhosis.
Sequential immunohistochimical study showed no change in control group (n=24). In 2 % formalin group (n=36), a mononuclear infiltrate inside and around the bile duct occured 7 days after injection and persisted until one year. At the later stages of the disease, Lymphocytes infiltrations were mainly T-cells. Ia antigene was expressed in biliary epithelium since the 7 th day.
Cholangiographic study showed normal aspect after hypertonic saline (n=11),and 0,5 % formalin (n=11). Strictures and dilatations of the biliary tree were observed in 10 rats after 2 % formalin injection (n=11). These strictures were confined to the ampula and the convergence in 4 rats, and located in the intra hepatic ducts in the others.
Our results suggest that 20% hypertonic saline and 0.5% formalin provide epithelium changes without sclerosis. 2% formalin provide sclerosing cholangitis Immunohisto-chimical study suggests a local cell mediated immune reponse in the pathogenesis of these lesions. 2 % formalin provide an experimental model of sclerosing cholangitis. Semeiotica Chirurgica, University of Padova, Italy. Severai Authors suggested the use of human fibrin sealants in pancreatic surgery to prevent fistuias. We performed a perspective randomized study incIuding 97 patients who underwent resective or derivative surgery, 34 femaIes and 63 maIes. 43 of them (Group A) had either the pancreatic anastomosis or the pancreatic resection seaied with fibrin glue during operation and 54 (Group B) had not. Twenty patients in group A and 26 patients in group B had an infiam matory pancreatic disease, whiie 23 patients in group A and 28 in group B had a pancreatic tumor (exocrine or endocrine). Sixteen patients in group A and 14 in group B underwent pancreaticoduodenectomy (PD); 16 (A) and 24 (B) had a pancratico-or cystoeunostomy (PJA); iO (A) and I3 (B) had a ieft pancreatectomy (LP) and i (A) and 3 (B) underwent a tumor excision (TE). The seaIant was made adding 500 U trombin to speed up soIidification and either the anasto motic edges or the pancreatic sutures were seaied with spray. AII the patients were treated and foiiowed-up by the same surgicaI staf The patients were randomized at the moment the surgicaI treatment was chosen, ividing the patients into 2 different Iists going to have either a resective or a derivative operation. We considered onIy radioiogicaIiy assessed fistuIas. After surgery weobserved an overall number of 12 fistuIas in 97 operations (12.7). Six fistulas were found in group A (13.9) and 6 in group B (11.1%). Five fistulas (16.1) occurred in patients with pancreatic cancer (3 gr. A and 2 group B), 6 (13.0) in patients with pancreatitis (3 in gr. A and 3 in group B), 1 occurred in a patient in group B with an endocrine tumor. According to the surgical procedure, we observed 5 fistulas (16.)in case of PD (4 in gr. A and 1 in gr. B), 5 (12.5) after PJA(2 in gr.A and 3 in gr.B), in 1 patient in group B after LP and in 1 patient in group B after TE. Our results do not show any statistically significant difference between the patients treated with fibrin sealant and the control group. We point out that the highest incidence of fistulas was observed in the patients with pancreatic cancer in group A (18.7) and in patients who underwent PD in group A (25.0).

Department of Surgery., Royal Postgraduate Medical
School, Hammersmith, I.bndon, W12 and rhe Thrombosis Research Institute, Chelsea, London, SW3, UK It has long been recognised that carcinoma of the pancreas is associated with a high incidence of thrombotic complications. Perhaps cancer cells generate procoagulant activity (PCA), which would not only predispose to thrombosis but also play an important role in mmour dissemination. Antithrombotic therapy at the time of implantation of certain experimental mmours will reduce metastatic potential. Since expression of tissue factor (TF) on the cell surfac may activate local coagulation, expression of TF is an important determinant of PCA.
We studied the PCA of 4 cell lines derived from human pancreatic adcnocarcinoma with normal and factor VII-deficient plasma to assess the role of TF in the generation of PCA. There is a great variation in the ability of human pancreatic carcinomas, of different origin, to generate thrombin, and this is particularly marked in TF dependent pathways. Should this ability correlate with biological behaviour, a new therapeutic pathway using specific antithrombins might become available.  HPB-unit, Academic Medical Center, Amsterdam The aim of this study was to evaluate the value of partial resection of the portal vein or superior mesenteric vein in patients undergoing pancreaticoduonectomy for carcinoma of the pancreactic head region. Patients and methods: Between 1983 and 1992 196 patients underwent pancreaticoduodenectomy for carcinoma of the pancreatic head region. Partial resection and reconstruction of the portal vein or the superior mesenteric vein was performed (end to end anastomosis or venous graft) in 20 pts in which infiltration of the portal vein or superior mesenteric vein was found peroperatively, in combination with subtotal pancreaticoduodenectomy (11 pts) and total pancreatectomy (9 pts). Results: In the group of patients with venous resection hospital mortality was 10% (2pts). Major complications were seen in !1 pts (55 %)" haemorrhage in 5 pts, intra-abdominal infections in 5 pts and mesenteric vein thrombosis in 1 pt.
Relaparatomy was necessary in 5 pts (25 %) and percutaneous drainage in 6 pts (30). In the group of patients with standard total pancreatectomy without venous resection (24 pts) hospital mortality was 16% (4 pts) and morbidity was 77% (19 pts). In the group of patients with subtotal pancreatectomy without venous resection (152 pts) hospital mortality was 5% and morbidity 40%. In the group of patients with venous resection, pathology showed a pancreatic carcinoma in 11 pts, an ampullary carcinoma in 4 pts and a distal bile duct carcinoma in 5 pts. Only 3 pts had a radical resection (tumour free margins).
One year survival was 26% and two year survival 11%, as compared to 69% and 50% resp. in the group of patients without venous resection. The tumour size and the differentiation grade were not significantly different from those in the group of patients without venous resection. .Conclusions" Partial resection of the portal vein or superior mesentedc vein in (sub)total pancreaticoduodenectomy for infiltrating carcinoma resulted in a low rate of radical resections and did not lead to improved survival. Thcrc arc several alternative methods to re-establishing continuity of the gastrointestinal tract following Billroth II conventional (B2PD) or Longmircreconstructed pylorus preserving pancrcatoduodcncctomy (L-PPPD). B2PD abolishes the postprandial gastrin response whereas the response after L-PPPD is similar to controls. In the present study a novel reconstruction after PPPD has been designed to separate the biliary and pancreatic secretions by restoring continuity in the Billroth I manner with the pancreatic remnant anastomoscd to a separate dcfunctioncd Roux loop (D-PPPD).
Gastrin pmol/L and CCK pmol/L responses wcrc measured in the fasting state and after a standard meal of 250mi 15% protein 30% fat 55% CHO with an energy value of 525I(J. Basal gastrin and CCI( wcrc similar (p < 0.I) in controls, B2PD and D-PPPD patients. In controls and D-PPPD but not in B2PD, postprandial gastrin rose at 10min from 8 + 1.9 pmol/L to 13 + 2.3 pmol/L, and from 7 + 2.1 pmol/L to 15 + 1.9 pmol/L, and remained elevated for 20 minutes. In the D-PPPD patients CCK levels remained low (0.2 +_ 0.09 pmol/L) after the test meal. In controls postprandial CCK rose from 0.5 + 0.1 pmol/L basal to 11 + 2.1 pmol/L and returned to basal at 60 min. In B2PD post-prandial CCK rose at a similar rate and remained elevated for 80 min.
CCI( is trophic to the pancreas. Abolition of the postprandial CCK response after D-PPPD removes this important potential stimulus for turnout rcgrowth in patients rcscctcd for cancer. The propensity for leakage and disruption at the site of the pancreaticojejunostomy is a major reason for morbidity and death after pancreaticoduodenal resection. The purpose of this study wa to evauate the roe of pancreaticogastrostomy as an aternative method of restoring pancreaticointestina continuity after pancreaticoduodenectomy. From January 1989 to December 1992, 32 patients have undergone pancreaticogastrostomy after pancreaticoduodenectomy at our institution. Tenty two patients were men, and ten women. The mean age as 60.5 years range 42-74 years Pancreaticoduodenectomy was performed for pancreatic carcinoma (16 patients) ampuary carcinoma 110 patients), duodena carcinoma (2 patients) and chronic pancreatits (4 patient There was one postoperative death, for an overa operative mortality rate of 3%. There was one pancreatic fistula 13%) which recovered ith further surgery. The average postoperative stay in the hospita was 14 days.
These results agree with data concerning pancreaticogastrostomy published in iterature 11). Trypsin neutraisation by gastric acidity may explain the very gow incidence of pancreatic fistuga reported. In ong term foow-up externa pancreatic insufficiency does not seem to occur. Radioogic and endoscopic examination of the pancreatic duct is easy. This method of reconstruction merits widespread utiisation due to tis simplicity and safety.
(1] DELCORE R. AND AL Pancreatogastrostomy, a safe drainage procedure after pancreatoduodenectomy. Delayed gastric emptying (DGE) has been reported as a frequent complication of PPPD, with an incidence of about 30 %. The cause of this phenomenon is probably multifactorial. We studied gastric myoelelectric activity, gastrointestinal symptoms and gastric emptying in the preoperative and early postoperative course in pancreatic and periampullary (PP) cancer patients to elucidate the possible role of gastric motor dysfuntion in the etiology of DGE in these patients.
G&,tric emptying (GE) of a nutrient-rich liquid meal was studied in 22 preoperative PP cancer patients with no duodenal obstruction as evidenced by endoscopy or radiography (13 ', 9 ; mean age 61 + 14 yrs) using Applied Potential Tomography, a technique measuring changes in resistivity in a thick cross-section through the abdomen at the level of the stomach. Myoelectric activity was measured concurrently by surface electrogastrography (EGG) during hour fasting and 1/ hours postprandially and analyzed by computerized power spectrum analysis. Thirteen of these patients were also studied within the first 2 weeks following PPPD. Also symptoms of nausea, vomiting, regurgitation, early satiety and fullness were scored on a scale from 0 to 3 both preoperatively and postoperatively. The same measurements were done in 25 sex and agematched healthy controls. Delayed gastric emptying was defined as t/ > 180 rain, being the mean + 2 SD of the t/ in these normal subjects.
Gastric emptying was delayed in 7 preoperative PP cancer patients (32 %). There was a significant negative correlation between g&tric emptying and pre-and postprandial mean gastric frequency in these patients. DGE occured in five of 13 postoperative patients after PPPD (38 %). The postprandial: fasting power ratio in the postoperative patients was significantly decreased compared to preoperative patients and controls (resp. 1.97; 4.92; 7.60, p < 0.05), indicating a decrease in gastric motor activity in these patients. Three of the postoperative patients with DGE also presented with DGE in the early postoperative period (60%). Also there was a clear association between postoperative DGE and the occurence of surgical (intraabdominal) complications. Gastric dysrhythmias were not observed in the pre -and postoperative period. No significant correlation could be demonstrated between DGE and symptoms.
Conclusions: 1) Preoperative DGE occurs rather frequently in PP cancer patients, which probably involves disease-related effects. 2) Gastric  yearsrange 21-74,F/MffiS/10 ). We also discuss which surgical approach to splanchnlcectomy should be preferred in the presence of tumors located on the head ,or body and tall of the pancreas. to 2.6 }. On the left both SN and SG are completely covered by superior edge of the head and body of the pancreas and on the left side their detection s rendered easy after splenopancreatIc moblllzatlon. The shape of the ganglions is varies from a thlck streak Intermlngled wlth swellings and slngle masses, rosary beads-llke. The gangllons are always firm, pearlcoloured, easily separated from lnfolandular nodes. The postganglonar fibers spread from GS, lke a shower, n the pancreas, and surround. In the chest, right and left SN get contributions of fibers from the sxth to the tenth thoracic ganglla.These fbers become a sngle bundle at a distance of 3.6 cm from the dlaphragm, always strictly thickened to the prevertebral fascla.From May, 1990 to June, 1992 we submitted 20 pts with unresectable pancreatic cancer to billary-enterlc by-pass plus blateral splanchnicetomy. All the pts had complalned severe upper abdominal paln.Postoperatve ortalty was nil at 1-month postoperative follow-up 95 % of pts achieved complete pan rellef.
Mean survival tSme was 7.9 months, and mean pan-free Interval was 5.7 months.Pan recurrence was manaued bY Dercutaneous cervical cordotony (1),cordotomy+narcotlc drugs (5), and narcotic drugs only( 12 .Two pts ded of progression of pancreatic ancer being rendered completely pan free. Two patient are alive without pan recurrence after 6 and 8 months respectively from the operation. Palliative surgical procedure offers considerable benefit for the patients with unresetable pancretlc cancer. Taking into account the above mentioned results we beleve that neoplasms of pancreatic head are the most suitable conditions to perform a surgical splanchnlcectomy by the right side through the hepatoduodenal legamentum,wlth a complete moblzatlon of IVC, whereas the retroduodenal-pancreatc approach seems to be more dlffcult. On the left the approach s particularly complex. The approach trough the , .or edge of the pancreas, suitable for pancreatc body or tal neoplasms, s complicated by the presence of the tumoral mass, as reported in our surgical seres. In presence of left neoplasms the transhatale approach by Dubos s more sutable,n spte of ts complexity. Recurrence of pan was treated by comblnaton therapy, such as percutaneous cervical cordotomy and/or narcotic drugs. Resection for pancreatic cancer is associated with a 5 year actuarial rate (YASR) of 5-15%. Adjuvant therapy may be useful but there has only been one study published. The North American Gastrointestinal Tumour Study Group (GITSG) phase III study reported an improvement following post-operative radiotherapy and 5-FU i.v. weekly in 20 patients with pancreatic cancer (43% 2 YASR vs 18% in control arm).
In order to ascertain the value of the GITSG protocol, this was repeated by members of the UK Pancreatic Cancer Trials Group. 27 patients with adenocarcinoma of the pancreas (N=23) or periampullary adenocarcinoma (N=4; one with resected solitary liver metastasis) have so far been recruited. There were 16 men and 11 women with a mean age of 59.4 (range 40-76) years.
A standard Kausch-Whipples operation was performed in 11 patients, 13 had a pylorus-preserving head resection, 2 had a total and 1 had a distal pancreatectomy. Lymph node metastases were present in 17 cases; the histological grade was 1=8, 2=11 and 3=8.
The radiotherapy was well tolerated with only one related complication There was no significant drug toxicity. t a median follow-up of 12 (range 3-36) months the 2 YASR was 30%; 14 patients were dead, 12 alive and 1 lost to follow-up.
These results indicate that post-operative radiochemotherapy is well tolerated. Gamma linolenie acid (GLA) is a eytotoxie agent for certain cancer cells. We have investigated the effects of GLA and its lithium salt (LiGLA) on hepatoma cell growth as regulated by hepatoeyte growth factor (HGF) and transforming growth factor beta (TGFI3).
The human hepatoma cell line, PLC/PRF/5, was cultured with or without eytokines, in the presence or absence of fatty acid (GLA and water soluble lithium salt GLA-LiGLA). The cell growth was quantified by a colodmetrie method and shown as percentage growth compared with cytokine and FA free control. Statistics is Student T test and significant level was taken at p<0.05 (*). Aim of the work we investigated whether the direct intratumoral transduction Of a suicide gene might induce the elimination of an established liver tumour. To specifically target the delivery and expression of the suicide gene into dividing cells, we used retroviral-mediated gene transfer and the herpes simplex virus type ] thymidine kinase conditional toxin (HSV|-TK). Non toxic nuc|eoside analogs such as ganciclovir are metabolized by HSVI-TK i nto compounds that are toxic for dividing cel | s.

Animals and methods
In 25 BDIX rats, a single liver tumour was induCed by injecti6n of DHDKI2 colon cancer cells under the liver capsule. At Day 5, tumour diameter ranged between 2 and 3 mm rats underwent an intratumoral injection of cells producing either HSVI-TK (treated group ; n=13) or (a nucleargalactosidase encoding marker gene) nls-lacZ (control group n=12) expressing recombinant retroviral particles. At Day lO, all the rats were given ganciclovir (150/mg/kg twice daily for 5 days). Autopsy was performed at Day 15. Results a dramatic reduction in tumour volume was noted in the rated compared to the control group 8.1 +_ 6.7 and 86.3 + 65.1 mm , respectively, p KO.O00|). In the treated animals, the esidual tumours were mostly made up of a massive fibrotic reaction with few or even no remaining tumoural cells.

Conclusion
Our study shows the regression of an established i-iver tuhiour after in vivo transfer of a suicide gene. This efficient therapeutical approach might be an ultimate treatment for disseminated liver metastases in humans, and could also be applied to the treatment of a large variety of solid tumours. Interleukin-3 (IL3) is a haematopoietic growth factor and has been used following radical chemotherapy for solid turnouts However there is a growing concern as to its possible stimulation of the malignant cells. We have investigated the effects of IL-3 on hepatoma cell growth. IL-3 showed a significant concentration dependent stimulation of the cell growth and the stimulation was observed from day 2 to day 7. To determine the possible relationship between IL-3 and other cytokines on cell growth, a variety of other cytokines were tested. The stimulatory effect of IL-3 was significantly enhanced by IL-4, IL-6, and IL-10, but was completely reversed by IL-8 and TGFII. There was no significant effect on IL-3 induced cell growth by IL-1, IL-2, IL-5, IL-7, and TNFa.
It is concluded that IL-3 is a hepatoma cell (PLC/PRF/5) growth up-regulator and therefore an important factor in controlling hepatoma growth. The liver constitutes the majority of reticuloendothelial system (RES) function, providing a barrier to the passage of intestinal bacteria and toxins to the systemic circulation. Intraperitoneal biomatedal promotes bacterial translocati0n from the gastrointestinal tract. However, the role of hepatic immune function against bacterial translocation is not defined. The purpose of the study was to investigate the role of hepatic function in the resistance against bacterial translocation after intrapedtone biomatedal implantation.
Meth0ds; Rubber drains, knitted dacron and silicone elastomer with areas of 3 and l0 em2, respectively, were implanted into the right-lower part of the abdominal cavity of Sprague-Dawley rats under aseptic condition. Hepatic retieuloendothelial function, epresse as the phagocytic index (k) and hepatic uptake of 125I-labelled E. oli was measured 1 day after implantation.

Results:
A significant elevation of the phagocytic index (k) was noted in groups with intrapedtoneal implantation as compared with sham operation. The hepatic uptake of 25I-labelled E. coli significantly increased after intraperitoneal implantation, of different biomatedals. Translocation of bacteria from the gastrointestinal tract was also observed. The increase in phagocytic index/hepatic uptake of 2SI-labelled E. coli and the incidence of bacterial translocation depended on the size of the biomatedal implanted. An inverse correlation existed between the hepatic uptake of 2SI-labelled bacteria and the incidence of bacterial translocation to the liver.

Conclusion:
Hepatic RF function increased after intrapefitone implantation of biomatefials, reflecting a enhanced phagocytic function of Kupffer cells. It appears that the liver is an important part of host defense in preventing bacterial translocation induced by intraperitoneal implantation of biomaterials. Surgery in patients with obstructive jaundice causes deleterious hemodynamic disturbances, possibly related to gut derived endotoxemia. This study evaluates the effects of preoperative binding of gut endotoxin by cholestyramine on hemodynamics with special reference to splanchnic blood flow. Male Wistar rats (n= 20 per group, weight 250-300 g.) received the endotoxin binder cholestyramine (CHOL, 150 mg/day) or normal saline (SAL) twice daily in the same volume. This treatment started on Day 1 and was continued until the end of the experiment. On Day 7 groups were randomized to receive a SHAM operation (SH) or bile duct ligation (BDL). This resulted in 4 groups (n 10 each) i.e. SH-SAL, SH-CHOL, BDL-SAL and BDL-CHOL. Subsequently rats were subjected to a surgical trauma by performing a xyphoidectomy on Day 20 and splanchnic organ blood flows were measured the following day using the radiolabeled microsphere technique.
A portion of the splanchnic organ blood flows is shown and expressed in mlmin "1 (mean +/-SEM). Portal blood flow was computed by the sum of the flow to the individual splanchnic organs. Surgery in saline treated BDL rats resulted in a significant decrease in portal venous blood flow. This was due to a decrease in flow to the pancreas, small intestine and colon. Stomach flow was increased. Cholestyramine pretreatment prevented the decrease in portal venous flow by increasing blood flow through spleen, pancreas, small intestine and colon.
It is concluded that preoperative treatment with the endotoxin binder cholestyramine prevents the fall in splanchnic blood flow following surgery in the bile duct ,gated rat. Endotoxin is important in the development of postoperative complications in the bile duct ,gated rat and the reduced splanchnic blood flow might enhance translocation of gut derived endotoxin. Portal blood supply may play some role in ischemic treatment of experimental liver tumours. However, normal liver parenchyma can not tolerate 60 min of repeated total devascularizations (occlusion of artery and portal vein). In this experiment portal deviation (portocaval shunt) was replaced with ligation of the portal branch supporting the tumour-beafing lobe. Twenty seven rats were subjected to sham-treatment (n=6); portal branch ligation (PBL) (n =7); 120 min of repeat dearterializations (n =7) and portal branch ligation (PBL) combined with 50 min of repeat dearterializations (n =7). The results showed that portal branch ligation alone did not alter the tumour growth compared with sham-treatment (p > 0.05).
However, 120 min of repeat dearterializations effectively decreased tumour growth (p<0.05 vs all the other groups). There was no growth delay following portal branch ligation in combination with 50 min of repeat dearterializations (p > 0.05). Significant atrophy of the corresponding lobe followed PBL regardless if it was combined with repeat dearterializations or not (p < 0.001). Aminotransferases remained normal after PBL with repeat dearterializations for 50 min (p > 0.05).
In conclusion, portal branch ligation did not retard tumour growth by its own, nor did it increase the antitumour effect following repeat dearterializations for 50 min. This study further confirmed that repeat dearterializations for 120 rain alone is to delay tumour growth. Phosphatidylcholine (PC) also decreased but without a significant difference compared with control (p > 0.05). The ratio of tumour/liver phospholipids was significantly reduced compared with control (p< 0.01) after a single dearterialization. Thus, dearterialization induced a fast degradation of tumour phospholipids without affecting the normal liver.
Repeat dearterializations for 2 hours during 5 days further reduced total tumour phospholipid. PC and PE were more significantly affected by repeat deartedalizations (p < 0.01 and p < 0.05 vs control respectively). The ratio PC/LPC (lysophosphatidycholine) dropped remarkably (p < 0.01 versus control) indicating relative accumulation of LPC.
The decrease of phospholipid was consistent with the growth delay of tumour undergoing repeat dearterializations. Phospholipid degradation was paralleled by a delay in tumour growth that was significantly retarded by repeat dearterializations (p < 0.01 vs control). Further, ASAT and ALAT remained normal even after repeat dearterializations for 5 days, which suggested that repeat dearterializations was selectively delivered to tumour tissue sparing normal liver cells. In conclusion, repeat dearterializations selectively induce liver tumour ischemia and efficiently retarded tumour growth without any visible damage to normal liver parenchyma. In the year 1992 we performed 151 laparoscopic cholecystectomies. After the optimistic start certain problems arose, all resulting from a lack of experience. Fortunately, all complications but one a hepatic duct injury were managed without major harm to the patient. Among our patients there were 120 women and 30 men, ranging in age from 13 to 82 years. The laparoscopic procedures were performed by four surgeons, alternating in the roles of the operator, the cameraman and the assistant. Conditions such as acute gallbladder infection and large stones, formerly constituting a strict contraindication, lost their significance with increasing number of cases. Among our first 50 patients there were six cases of conversion; later on no conversions were necessary.
The preoperative assessment, including the patient's history, laboratory parameters and high-quality US scans, was considered adequate, offering reasonable guarantee that no significant bile duct pathology would be overlooked. Intraoperative laparoscopic cholangiography was performed without problems in only four cases. The mean operative time was 65 minutes and the mean hospital stay 3.5 days; the patients returned to work in seven to ten days.
Conclusions: For the patient's benefit, surgeons are sacrificing their threedimensional sight and tactile capabilities. A one-handed surgeon with only one eye left would still be able to perform a laparoscopic operation. In the hands of a skilled and experienced "laparosurgeon" the method is perfectly safe. of these was referred for definitive treaent and had a HJ, but 2 initially underwent unsuccesful repair over a T-tube and ultimately had HJ. Six pients developed a common hepatic duct stricture (with a localised bile leak in 2). Five of these underwent successful HJ" the sixth was referred with a complex hilar stricture following an HJ and was treated by percutaneous dilatation. The median time from LC to the definitive surgical procedure was 18 weeks (range: 0 52). None of t patients had operative cholangiography and most BDI were not recognized immediately. The mechanisms of these injuries were not always clear at reoperation.
The inherent risk of BDI may be greater in LC than in OC, but this may be further increased by distortion of the anatomy, excessive use of diathermy or inexperience. The problem of training surgeons in laparoscopic procedures persists and should be urgently addressed. In a previous study(), we have shown that: (i) cholesterol stones can recur after postcholecystectomy stones; (ii) in patients with associated jaundice and pancreatitis, the removal of cholecystectomy and initially form in a long cystic remnant (LCR) (11 of 63 patients with the LCR determined the complete and persistent relief of symptomsk On the other hand: (i) the LCR itself rarely causes symptoms. Therefore, the excision of a LCR without associated stones will not eliminate postcholecystectomy symptoms in the great majority of patients; and (ii) the exact percentage of patients with LCR who are going to develop non brown gallstones many years after cholecystectomy is unpredictable, even if it is not negligible. Concerning stone recurrence, LCR can be defined as "a cavity lateral to the common duct, lined by gallbladder mucosa, acting as a minigallbladder, regardless of its length, ranging in our series from 0.8 to 4.5 cm".
After the wide diffusion of laparoscopic cholecystectomy, which deliberately leaves a long cystic stump, it is presumed that also the number of postcholecystectomy stones related to LCR is going to increase, in the long term period. Since these stones are mostly cholesterol, possible prophylaxis using bile acid therapy could be indicated in selected cases. Therefore, a double blind trial with bile acid therapy was undertaken. Twelve patients with a cystic stmnp longer than 15 mm after laparoscopic cholecystectomy were enrolled. Patients with a presumed residual LCR were selected among those showing a particularly long and tortuous cystic duct at cholangiography. In particular, in 2 patients the cystic duct had a very low confluence, close to the ampulla of Vater. One of the last 2 patients plus 5 of the remaining 10 with LCR were treated by tauroursodeoxycholic acid (7 mg/Kg/die) to prevent cholesterol stone reformation. The remaining 6 patients didn't receive any treatment and were used as control. Ultrasound examination was scheduled every year and cholangiography every 3 years. At the moment, after a 6-month-follow-up, no patients in either group developed symptoms or gallstones.  (Table)" METHODS. Fifty-seven well nourished patients candidates to cholecystectomy for uncomplicated cholelithiasis, without indication to common bile duct exploration, were prospectively randomized into two groups undergoing either LC (n=30) or OC (n=27), using the same type of general anesthesia. In each patient blood samples were drawn preoperatively and at various time intervals after surgery, to determine the concentration of lymphocyte subsets CD3, CD4, CD8, OKDR and the plasma levels of C-reactive protein (CRP), cortisol and prolactin.
RESULTS. The duration of operation and of general anesthesia was similar in the two patient groups. The postoperative hospital stay was shorter after LC than after OC (3.1(0.5) vs. 7.1(1.6) days; P<0.001). After OC a significantly greater (P<0.05) depression of blood CD3 and OKDR counts was found, as compared to LC. Moreover, OC patients showed a significantly greater postoperative acute phase increase of plasma CRP (P<0.001), of cortisol (P<0.05) and of prolactin blood levels (P<0.001).
CONCLUSIONS. The significantly lower acute phase responses of immunologic and metabolic parameters observed after LC support the concept that the laparoscopic procedure is less traumatic.  (180 kg). All patients have billary symptoms. Three pts were staged ASA IV and 11 staged III. This staging was due to cardiovascular or pulmonary failure for 14 pts and for morbid obesity for 1 pt. During the procedure radial artery catheterisation provided the arterial pressure values (systolic, diastolic, mean) and blood samples for blood .gases and pH measurement. Continuous capnography was made. Cardiac output, pulmonary artery_ and wedge pressure were provided by a Swan Ganz catheter witch also permitted the determination of systemic and pulmonary vascular resistances. In all pts anesthesia was Induced using etomidate (0,3mg/k.g) and fentanyl (7 microg./Kg) Atracurium (0,3 .mg/kg) provided myoresolution. Maintenance of anesthesia was realized with fentanyl,atracurium and 1,5 isoflurane in a mixture of nitrous oxyde and oxygen (FIO2:0,5).This protocol was accepted by the ethical board .The pneumoperitoneum pressure was always inferior to 12 mm of Hg. In 8 cases there was a cholecystitis, in 6 cases the gallblader was not inflammatory and in 1 case there was an empyerna of the gallbladder. No conversion .in laparotomy was necessary .The mean duration of PCC .was 76,3 +/-24,4 mn (52-150).

Results
Tumor hypervascularization including pulsatile signals at the periphery, in the centre of the tumor or both was observed in: 2 AFP negative of 4 cases with Hepatocellular caminoma (HCC), 4 of 4 with Cholangiocellular caminoma, 3 of 8 with giant hemangioma, 7 of 7 with Focal Nodular Hyperplasia, 3 of 4 with adenoma and furthermore in one neumendocrine tumor. The tumor was hypovascular in 2 patients with HCC one with status after chemoembolization and one with multiple intrahepatic tumor recurrence on underlying posthepatitic cirrhosis after liver transplantation 6 years before. Furthermore, the tumor was hypo-or even avascular in 5 patients with hemangioma and in one with adenomatous hyperplasia.

Conclusion
With regard to the vascularization pattern, yield of Color Doppler sonography is low for differential diagnosis and prediction of tumor dignity. However, valuable information about tumor extension can be obtained in central tumor lesions close to hilar structures. Early recurrence is frequent after surgical resection of hepatocellular carcinoma. The purpose of this work was to assess the value of postoperative adjuvant regional chemotherapy. Cisplatin (2 mg/kg bw) diluted in 10 ml of iodized oil (Lipiodol, Lab. Guerbet) was administered after selective catheterization of the hepatic artery above gastroduodenal artery one month after resection and whenever possible 5 and 9 months postoperatively. Between April 1991 and December 1992, 24 patients had liver resection for treatment of a hepatocellular carcinoma. Thirteen patients were excluded from the study for one of the following reasons: residual cancer nodules after operation (4 patients), markedly impaired liver function or ascites (5 patients), spontaneous or surgical total portal diversion (2 patients), failure in selective catheterization of hepatic artery (2 patients). Eleven patients received the treatment: 6 patients had three courses of chemotherapy, two patients had 2 courses and three patients had one course. Side effects of the treatment were minimal and the mean hospital stay was 3 days. Two patients had prolonged epigastric pain due to gastritis. Recurrence was looked for in each patient by ultrasonography, CT scan and measurement of serum (fetoprotein every 4 months. One patient died on the fourth postoperative month from spontaneous bacterial peritonitis. All other patients are alive. None of them demonstrated recurrence with a mean follow-up of 10.7 months (range 4 to 18 months). Although our results are preliminary, they suggest that adjuvant regional chemotherapy using cisplatin mixed in iodized oil might be efficient in the prevention of early recurrence after resection of hepatocellular carcinoma. Recurrence occurred in 38 pts. (33.6%). The 3 and 5 year actuarial survival rates for pts with follow-up > 30 days has been 60.1% and 36.3% respectively. Univariate analysis of factors influencing survival has shown better results for pts. 1) without symptoms; 2) with AFP level < 20 ng/dl; 3) who did not receive blood transfusion during surgery; 4) who had operation shorter then 3 hours; 5) without satellite nodules around the main tumor; 6) without microvascular thrombosis and 7) when the tumor was completely capsulated. On the other hand Cox's proportional hazards regression model has shown that only 1) the presence of a complete peritumoral capsule, 2) the smaller tumor diameter and 3) the absence of microvascular thrombosis were independent factors predicting statistically better survival (p<.01). The 3 and 5 year actuarial survival rates for the group of 21 patients with a11 these 3 variables have both been 79.1%. CONCLUSIONS Surgery can be proposed for cirrhotics with small, encapsulated HCCs and good hepatic function. Intraoperative US is an essential tool to avoid unsuccessful operations and to guide limited anatomical resections. Long term survival could be expected for selected patients. Our results suggest (1) a reduction in the incidence of HA, possibly due to the use of low-estrogen-content OCs in recent years, (b) an apparent increase in the diagnosis incidence of FNH, probably due to the diffusion of ultrasonography, (c) that a hypervascular tumor with a visible central element in an asymptomatic young woman is almost certainly a FNH and (d) that enhanced MRI and biliary scintigraphy are the best imaging techniques for the diagnosis of FNH. Twenty percent of liver resections for hepatocellular carcinoma (HCC) are done in patients with a non cirrhotic liver. The purpose of this work was to assess the results of liver resection for HCC in 40 patients with a normal liver. Mean age (27 males and 13 females) was 56 years (range" 13 to 75 years). In 85 % of the patients there were symptoms" pain, fever, or weight loss. None of the patients had liver failure or portal hypertension. One patient had jaundice resulting from invasion of bile duct. There were 14 right, 10 extended right, 6 left and 2 extended left hepatectomies, and 8 segmental liver resections. One patient had resection of the bile duct. A vena caval neoplastic thrombus was removed in 1 patient. Liver resection was curative in 35 patients and palliative in 5. Mean tumor size was 123 + 75 mm (range" 30 to 280 mm). The tumor was encapsulated in 58 % of patients. In 64 % of those, there was capsular invasion. Satellite nodules were present in 80 % and invasion of distal portal branches in 76 % of patients. There was no operative mortality. Five patients (12 %) had a major postoperative complication'intestinal obstruction in 1 and deep jaundice in 4, which subsided in 3. In the last patient jaundice persisted leading to major liver failure. On the 17th postoperative month this patient is awaiting liver transplantation. One-, two-, and three-year survival rates were respectively 80, 72, and 35 %. Tumor recurrence occurred in 28 patients (72 %) and was confined to the liver in only 13. These results suggest that the prognosis of HCC in patients with a non cirrhotic liver undergoing resection is approximately the same as that of patients with cirrhosis. The poor long-term survival might result from the marked spreading of the cancer within the liver due to late diagnosis in a population with no survey. There were 7(5.4%) intraoperative complications. In 2 instances of fight hepatectomies extended to the segments IV and I we have injured the left hepatic duct. In all 2 cases they are sutured using a T tube protection. In 3 patients with large mmours adherent to the cava vein we caused a small injury of the latter: total hepatic vascular exclusion was necessary for 5 minutes in 2 patients. A partial improper closure of remnant left hepatic vein was made in the sixth patient, who had a fight hepatectomy extended to the segment IV. An end-to-end suture was performed in total hepatic vascular exclusion. In the last patient, who had a fight hepatectomy and a portal resection for hilar cancer, we had a torsion of an end-to-end portal anastomosis and we performed a reanastomosis. The median blood transfusion requirement was 2 units (range: 0-7); 37 patients did not require blood. The median operative time was 288 minutes (range: 150-600). Intraoperative complications were strictly related to the tumour size: 6 out of 7 patients had a tumour diameter of 10 to 30 cm. (average: 18.1 cm) interpositlon-graft portosystemic shunts maintain that the incidence of encephalopathy post-total portosystemlc shunt is secondary to loss of hepatic portal flow across a low pressure gradient anastomosis. This study assessed and correlated the incidence of post-op encephalopathy in 72 patients with alcoholic cirrhosis, portal hypertension, and variceal hemorrhage [Childs' class A (6%), B (40%), C (54%) undergoing a side-to-side portacaval anastomosis (PSmm) who had had portal hemodynamlc studies pre-op, intra-op, and post-op. Sixteen patients (22%) developed clinlcally-evident post-op encephalopathy: 9/16 were not encephalopathlc pre-op, 6/16 occurred early (<30 days), 10/16 were late (sepsis 4, dietary indiscretion 3, multiple and incapacitating 3), 4/16 died post-op, there was no difference in measured portal hemodynamic parameters in these 16 patients compared to 56 patients without encephalopathy (Table). Encephalopathic patients had worse hepatic function (Childs' class C 70% vs 46%) and higher incidence of emergency shunts (38% vs 9%). Encephalopathy occurs post large diameter portacaval anastomosis as a result of failure to acutely control gastroesophageal bleeding by nonoperative means in patients with poor hepatic reserve and is not secondary to the size of the anastomosis or the post-shunt pressure gradient across the anastomosis. A large diameter (25ram) side-to-side portacaval shunt with an 8% mortality, a 22% encephalopathy rate (new 12%, incapacitating 4%), and no recurrent variceal bleeding remains the "gold standard" for the treatment of variceal hemor::nage in patients with alcoholic cirrhosis.
Pressures(mmH20),XSEM Material and methods: Standard end-to-side portacaval shunt (n=29) and shamopcration (n=29) was performed under ether anesthesia. Four weeks after surgery rats were sacrificed and segments of proximal and distal small bowel was taken in a standardized fashion for light and electron microscopy. The specimens were examined blindly. Villus height, villus to crypt ratio and villus size index were assessed on light microscopy. Microvillar height and microvillar count was analyzed by clcctronmicroscopy.
Results: No differences could be detected in villus to crypt ratio, villus size index and villus height between shunted rats and rats with sham operation. The vilus size index was high in the proximal small bowel in both groups and gradually decreased along the length of the intestine. On electron microscopy no differences could be detected in microvillar count per unit length of the brush border between shunted and sham rats. On the other hand microvilli of enterocytes from control rats were significantly taller compaxed to those in the shunted groups.
Conclusion: In contrast to other studies we have not found any changes of the small intestinal mucosa after .shunting detectable at light microscopy, but a significant reduction of the height of microvilli on electron microscopy. The role of portal venous pressure in the regulation of mucosal absorptive surface by altering the height of microvilli needs to be studied further. Decompression of the liver by either portacaval or mesocaval shunts for the Budd-Chiari syndrome complicated by inferior vena cava occlusion is impossible either due to hypertrophy of the caudate lobe or prone to failure due to high IVC pressure, as found during operation in our patients.
The outcome of splenopneumopexy is also poor as indicated by 2 patients in this series (one referred for mesoatrial shunt due to failure of this procedure performed elsewhere).
Six patients with adequate collateral circulation between the hepatic veins and IVC have had cavoatrial shunts.
Seven patients with adequate collateral circulation of the occluded IVC have had mesoatrial shunts. One patient with severe venous hypertension of the lower extremities (varicosity, edema, pigmentation, and venous ulcer) had combined mesocaval and cavoatrial shunts. All of these shunts were constructed using a 14 or 16 mm ring reinforced polytetrafluoroethylene graft.
Fourteen patients treated by shunts recovered well with resolution of ascites, diminution in liver size, and improvement in IVC pressure gradients.
One patient complicated by post necrotic cirrhosis died of hepatorenal syndrome after a cavoatrial shunt. The standard surgical treatment of Budd-Chiari syndrome (BCS) is still a shunt procedure between the portal and the caval system. Depending on whether or not the inferior vena cava is obstructed, the presence of a negative or weakly positive porto-inferior vena cava (IVC) gradient precludes a shunt bypass toward the IVC system. In that feature, a shunt bypass toward the superior vena cava system is needed. Between 1/1973 and 12/1992, we performed 71 shunts on 62 patients. In 37 cases (34 patients), portal flow was shunted toward the IVC. In 34 cases (26 patients), the portal flow was shunted toward the SVC. From 111981 to 12/1986, 17 patients with a BCS and an inferior vena caval obstruction were treated by an intrapericardial shunt (IPS) (i.e. mesoatrial shunt). From 1/83 to 12/92, 17 patients were treated by an extrapericardial shunt (EPS) avoiding opening the pericardium. Since 1/87, 14 of these patients were treated by two new procedures, namely a mesoinnominate (MI) or a meso-extrapericardial superior vena cava shunt. All patients had pre and intraoperative liver biopsies (LB). LB were classified as follows centrilobular necrosis (CLN) alone, CLN and severe fibrosis (F), F alone. There were 16 CLN, 28 CLN and F, 27 F.12 patients had a preoperative acute renal failure (POARF) defined by a blood creatinine level > 120 imol/l. Of these 12 patients, 9 had F.

Results.
Overall results :The overall mortality was 29 %. 9 of the 12 patients with POARF died (75%). 8 of these 9 patients had F on LB (90 %) 1 patient over 9 with POARF and F survived (10%). The two patients with POARF and no F survived. Statistical analysis showed the following results POARF vs no POARF (p<0.001) F and POARF vs F and no POARF (p<0.001).
BCS with free IVC Total early mortality was 27 % (9/33) 1 intraoperative death, 4 septic shocks, 3 upper digestive bleedings, 1 pulmonary embolism. 4 of the 6 patients with POARF died. All had fibrosis on LB. 2 deaths occurred during the last 10 years and none during the last 5 years. Follow-up ranges from 1 to 15 years. One patient died 5 years later of unknown mason. Actuarial survival rate is 73% at I year, 73% at 5 years, 66% at 10 yearn. BCS with IVC obstruction IP_P 70% of the patients had postoperative ascites. Postoperative mortality was 57%. Of note, between the second and the fourth postoperative month, 4 patients (23%) had a pericarditis (2 acute and 2 constrictive) leading to a pericardial tamponade. 3 of them died even though a pericardectomy was performed. EPS Of the 14 patients treated since 1987, 1 patient died of a postoperative septic shock.
Overall mortality was 7%. No pericarditis occurred. According to previous studies the sheep liver is most suitable for segmental resections. In trial I (3 groups of 10 sheep) either segment II,III or IV was removed. Resection (SR) was performed after intraoperative ultrasound guided localisation of the corresponding segmental portal branch, its ligation after transparenchymal dissection and staining of the segment by injection of methylenblue. In trial II (2 groups of 10 sheep) a traditional resection (TR) of segment III or a sham-operation (laparotomy including 200 ccm bloodloss) were performed. Dissection of liver tissue was carried out with an ultrasonic aspirator (Sonoca, Quickbom). Precision of segmentectomy was evaluated by corrosion preparation of the liver. Clinical profile was assessed by intraoperative bloodloss and changes of liver function tests (postop. day 1-5). Results: Overall, 36 corrosion preparation were technically sufficient for assessment. In SR anatomically precise resections could be achieved for segment II and III in 66 % each and for segment IV in 50%. In TR only one resection was anatomically precise (p<0.01, X2-test).
Bloodloss for resection of segment III was nearly the same in SR (mean and SD: 148 ccm + 61) and TR (140 ccm _.+ 77). Liver function tests did not differ significantly between SR and TR.

Conclusion:
The new technique pro.red to be a safe and reliable surgical procedure in anatomically precise liver segmentectomy which might have advantages in patients at increased perioperative risk. Auxiliary partial heterotopic liver transplantation (APHLT) could be an attractive treatment for fulminant hepatic failure in which recovery of host liver might be expected. We studied the functional relation between auxiliary heterotopic partial liver graft and host liver in the presence of artificially created portal hypertension in pigs.
16 pigs underwent hepatic artery ligation and APHLT, and were randomly allated to three groups as follows; group A(n=5)" no treatment of host lrtal vein, group B(n=6): banding of host portal vein to make the host portal pressure higher than that of graft by 2 mmHg, group C(n=5)" banding of host portal vein to make the host portal pressure same as that of graft. Postoperative immunosupression was done with cyclosporin, azathioprine and steroids.
APHLT could be a good and acceptable treatment for temporary support in fulminanthepatic failure. However, itshould be mandatory to band the host portal vein to make the host portal pressure same as that of graft in cases without portal hylrtension. Bacterial infectious complications, including intraabdominal sepsis and bactcrcmia, following major liver resection can at least partly bc attributed to translocation of enteric bacteria. Attempts to prevent or treat such infections by use of antibiotics may instead result in colonization and/or overgrowth of surviving microbes. In the present study, the effect of enteric administration of phospholipids (phosphatidylcholinc and phosphatidylinositol) on the prevention of enteric bacterial translocation, induced by subtotal liver resection in the rat was evaluated. 90 % hepatectomy was performed in male Sprague-Dawley rats. The animals were allocated to receive saline, phosphatidylcholine or phosphatidylinositol prior to operation. Enteric bacterial translocation, intestinal mucosal mass and enterocyte protein content were determined.
The incidence of bacterial translocation significantly increased 2 and 4 h following 90 % hepatectomy, as compared with sham operated animals. Enteric administration of phospholipids, however, significantly reduced the incidence of bacterial translocation afte 90 % hepatectomy. Phospholipid treatment also prevented the otherwise occurring postoperative decrease in intestinal mucosal mass and enterocyte protein content.
Enteral administration of phospholipids thus seems to protect against translocation of enteric bacteria and prevent from the decrease in intestinal mucosal mass and enterocyte protein content following subtotal hepatectomy in the rat. The Uni,ersity of Ankara, Ankara, Turkey Despite improvement in operative techniques, better pre-operative evaluation and post operative care and uses of new antibiotics, mortality and morbidity due to infections and endotoxemia in obstructive jaundice still remain high.
In obstructive jaundice changes in bacterial flora occur because of absence of bile acids in gastro-intestinal (GI) tract and there is an increase in colonization of especially gram negative bacteria. Bacterial translocation (BT) may occur even though there is no damage to the intestinal wall.
In this study, effect of lactulose which decreases endotoxin levels and its absorption on BT was sought. Four groups of rats were studied; control (n: 10), sham ligation (n:20), bile duct ligation (n:19), bile duct ligation and oral lactulose (n'20). In the control group during the first laparotomy specimens from mesenteric lymph node, liver and caecum were taken for culture. In the other groups the specimens were taken 14 days after the operation.
In the control group, no bacterial translocation was observed. In the sham ligation, bile duct ligation and bile duct ligation with lactulose administered, group BT rats were %5, %36 and % 10 subsequently.
In the group whose bile ducts were ligated, there was an increase in E. coli and proteus colonization in the caecum. On the other hand in the lactulose administered group colonization of E. coli and proteus was found to be decreased.
Lactulose which decreases the endotoxemia exerts its effect by regulating the flora in the GT tract. The effect of lactulose on flora is through lowering pH in the caecum. The concept of endotoxin-mediated rather than direct liver injury in biliary obstruction was investigated using the experimental rat model of bile duct ligation (BDL) and small bowel bacterial overgrowth (SBBO). Small identical doses of intravenous endotoxin (LPS) were administered to sham operated rats as well as rats with BDL. LPS caused a significantly more severe liver injury in the BDL group, determined by the histologic score of the liver damage and serum gamma-glutamyl transferase, C-reactive protein and plasma LPS levels, suggesting the possible contribution of LPS in this type of liver damage within the framework of an altered Kupffer cellmacrophage system. After sensitization of the liver to minute amounts of LPS was documented, the possible role of intestinal endotoxins in precipitating the liver injury triggered by BDL was tested. In this respect, we surgically created jejunal self-filling blind loops (SFBL), which were known to result in SBBO and therefore increased intestinal endotoxin pool. When BDL was combined with SFBL, once again a profound liver injury was observed in contrast with the control group of rats with BDL+self-emptying blind loops (SEBL). Turnout necrosis factor (TNF) is an important mediator in inflammation, but excessive production may cause host damage. TNF inhibitor (TNF-inh) is a soluble protein which block TNF actitivity. We have investigated the monocyte production of TNF and blood levels of TNF-inh in patients with obstructive jaundice.
29 jaundiced patients and 23 controls were studied. Monocyte TNF production (LPS stimulated) was measured by the L929 bioassay and shown as U/ml. TNF-inh was assayed using a TNF cytotoxic inhibition assay I and shown as percentage inhibition of TNF activity.
Monocytes from jaundiced patients showed greatly increased TNF levels and also increased TNF-inhibitor levels compared with controls. TNF/TNF-inh ratios, which reflect the balance of TNF and its inhibitor, are shown as follows: Control Jaundiced All We conclude that jaundiced patients have greatly increased monocyte TNF production and also have increased levels of TNF inhibitor. The regulation of the balance between TNF and its inhibitor may be an important aspect to influence patients prognosis. Two groups of PPVL and cirrhotic Wistar rats were studied (n = 20 and 12 respectively). Half the rats in each group received a 20 rrn infusion of 8-OR (0.043 U/hour) whilst the remainder received saline (control groups). Pulse, mean arterial pressure (MAP) and portal pressure (PP) were monitored continuously. Cardiac output (CO) and organ blood flow were measured prior to and after infusion by a dual mierosphere technique. Intrarenal shunting was determined by a renal/pulmonary passing fraction method.
In conclusion 8-OV improves the hyperdynamic state found in PPVL and cirrhotic rats. Renal blood flow is preserved even though the cardiac output falls and intrarenal blood shunting in cirrhosis is reduced. These effects explain in part its beneficial effect on renal function in the hepatorenal syndrome. One of the most important prognostic factors determining survival among patients undergoing hepatic resection for colorectal metastases are the number of metastatic deposits in the liver. Since a number of more than 3 metastases may exclude a patient from resection, precise evaluation is essential to prevent needless exploration.
A prospective evaluation of the liver by preoperative ultrasound (US), conventional computed tomography (CT), and continous CT angiography (CCTA) was performed in 60 patients with and without metastases. All patients had a history of colorectal carcinoma and all underwent laparotomy. The standard of reference were the findings at laparotomy: palpation of the liver and intraoperative ultrasonography (IOUS). The imaging techniques were assessed independently of each other. One hundred and five liver metastases were identified in 37 patients; 42 metastases were less than 1 cm in diameter. Twenty-three patients had no metastases. CCTA had a high sensitivity of 94% (99 lesions identified), in contrast to US (48%, 50 lesions identified) and conventional CT (52%, 55 lesions identified). The superiority of CCTA is also manifest in lesions under 1 cm in diameter. However, the high sensitivity is accompanied by a high-false positive rate particularly due to variations in the perfusion of normal liver parenchyma. Overall, CCTA had the highest accuracy (74 %)compared to US and CT (both 57%).
The data indicate that preoperative US and conventional CT have low sensitivity in the detection of liver metastases and that CCTA can supplement information, although the low specificity of the technique makes its application difficult. Possibly, experience with image interpretation may allow the interpreter to accurately predict the likelihood that a specific lesion represent metastatic disease. In addition, combination with other imaging techniques may increase the accuracy. Concltsion: Although CCTA seems to be superior to other preoperative imaging techniques the too low specificity will hamper its routine application in patients with hepatic metastases from colorectal carcinoma. The use of mdiolabelled monoclonal antibodies to tumor-associated antigens in tumour diagnosis and detection is worldwide accepted. Recently a new tecnique has been proposed, based on the development of a hand-held gamma detecting probe for intraoperative use. We report our experience using radioimmunoguided surgery (RIGS) for resection of liver metastases from colorectal carcinoma. Since 1989, 54 patients with primary or recurrent colorectal carcinoma received preoperative intravenous injection of MAb B72.3, a mudne IgG-1 that reacts with the high molecular weight glycoprotein antigen TAG 72 associated to colorectal cancer. It was radiolabelled with I-125. A total of 24 patients with liver metastases (13 synchronous and 11 metachronous) were considered eligible for resection at preoperative diagnosis. At surgery the whole abdominal cavity and its viscuses were explored by the probe, in order to confirm lesions preoperatively assessed and detect possible occult tumour deposits. Intraoperative ultrasonografic scans were routinely performed and compared. Probe assisted liver resection permitted to locate accurately liver metastases, to exclude lymphnode involvement, to delineate margins before resection, to control margins during non anatomical resection and to verify clear margins after resection. Five of 24 patients (19%) were found to have a large bilateral metastatic spread not previously detected; major procedures were avoided and a catheter for locoregional chemotherapy was placed in the hepatic artery via the gastroduodenal artery. Thirteen of 20 (65%) synchronous metastases and 20 of 25 (80%) metachronous metastases were correctly localized by the probe. Twelve metastases (7 synchronous and 5 metachronous) were unrecognized. Factors influencing MAb binding to colorectal cancer cells (MAb pharmacokinetics and physiological factors within the tumor) negatively affect the radiolabelling of turnout metastases. The smaller lesion identified was 5 mm in diameter. Compared to intraopemtive US-scan RIGS shows a lower sensitivity (75%) and a higher specificity (100%  1, 1980 andDec. 31, 1990 data from 1099 patients with colorectal liver metastases were recorded at the University Hospital Erlangen. Excluding all patients, who received any kind of treatment, 489 patients remained for the univariate (log-rank) and multivariate (BMDP 2L) analysis of the determinants of the natural history of colorectal liver metastases. All patients were followed up to June 1, 1992 or death. At the closing date of the study only 11 patients were still alive.
Univadate analysis yielded the following highly significant factors: mesenteric lymph node involvement (MLNA)0 grading of the primary tumor, treatment of the primary tumor, radicality at the primary tumor site, hepatomegaly, percent liver volume replaced by tumor (%-LVRT), number of liver metastases, distribution of liver metastases, alkaline phosphatase, LDH, WBC-counting, CEA, Karnofsky-index and extrahepatic tumor. With multivariate stepwise regression analysis (Cox) 6 independent significant factors were selected" %-LVRT, grading of the primary tumor, MLNA, extrahepatic tumor and diameter of metastases. Subsequent combination of the multivariate significant factors resulted in a factor adapted prognostic assesment (FAPA). This prognostic "tree" demonstrates the heterogenity of this patient sample with median survival times ranging from 3.1 months to 20.3 months, dependent on the presence or absence of the multivariate, relevant factors.
Because of the extreme differences between the group consisting of less than 25%-LVRT and the group with more than 25%-LVRT we also carried out a multivariate stepwise regression 1analysis, selectively for each of these groups. For patients with < 25% LVRT mesenteric lymph node involvement was replaced by LDH as a significant variable allowing a more adequate definition of prognosis. In patients with > 25 %-LVRT only grading of the primary tumor and diameter of metastases were proved independently significant.
In view of the lack of detailed knowledge on the natural history of colorectal liver metastases the prognostic "tree" presented may improve the assesment of palliative therapeutic approaches, with respect to prognosis compared to the natural history of colorectal liver metastases. METHODS The patients were considered suitable for treatment if their lesion(s) was >5cm and <15cms in size, <10 in number and accessible to percutaneous puncture. A 19 gauge needle was inserted into the centre of the tumour under ultrasound control, a naked laser fibre was then inserted into the core and the needle withdrawn. Laser therapy was then administered by continuous firing until the lesion became echogenic. Patients were assessed clinically and by ultrasound imaging during and for the first 24 hours after therapy and again at 2 weeks. CT was performed at 8 and 24 weeks. RESULTS A mean 5000 Watts (range 1449-9449) dose was administered per session and each lesion received between 1 and 3 treatments (mean 1.8).
Median follow up was 31 weeks (range 7-46). There were 2 deaths during this time and 1 patient refused follow up. Six patients had completed 6 month follow up, in 3 the size of the tumour(s) decreased but 1 developed extrahepatic metastatic disease and in the other 3 progression of local disease was apparent. Two specimens (1 post mortem and 1 resection) were examined. Treated areas showed more haemorrhage and possibly greater necrosis than untreated areas. Seven patients required opiate analgesia post procedure for 12 hours and 3 patients required extended hospital stay for pain control. Other side effects included macroscopic haematuria (2), massive abdominal wall bruising (1), biliary leak (1), jaundice (1). CONCLUSIONS This study indicates that ILH is a feasible option for the treatment of hepatic malignacy. Cell necrosis and decreased tumour mass can be achieved despite the significant procedure related morbidity and therapy can be easily monitored by ultrasound. The value of ILH as an adjuvant to other therapeutic options in the management of patients with hepatic malignancy is worthy of study in a randomised clinical trial setting. With recent advances in liver surgery, liver resections with curative intent are being performed at a growing rate. However, there is a group of patients with lesions at the confluence of the hepatic veins who pose a problem for the hepatic surgeon. Extended resections for these lesions have a highr morbidity and mortality than lesser procedures (segmental resection). The use of the transverse hepatic plane to resect $5 and S6 has been previously reported. This surgically created plane can also be used to resect those lesions at the hepatic vein confluence. We report two cases of right and middle hepatic vein sacrifice with segmental liver resection of S7, S8, and S4A (cranial transverse hepatectomy).
In the last three years and seventy liver resections, our institution has seen several challenging lesions at the hepatic vein confluence. Two patients underwent cranial transverse hepatectomy for metastatic colon cancer. Each patient had a solitary lesion located at the junction of the IVC, RHV, aD MHV. There were no intraoperative or postoperative complications. One patient received no blood transfusion and the other 6 units. Discharge was at 14 and 12 days postoperatively. Peak bilirubin was 3.7 and 3.3. Bilirubin at POD 10 was 2.9 and I .6. Both patients are alive at 10 and 22 months.
Cranial transverse hepatectomy is a safe and useful segmental resection for lesions located at the hepatic vein confluence. The ability of the liver to develop collateral venous drainage to the lower segments (S5, S6, and S4B) is remarkable. Histologically, a pancreatic cystic tumour was found in 24/156 patients (15%), including 3 microcystic adenomas, 9 cystadenomas, 10 cystadenocarcinomas, 1 solid and papillary epithelial neoplasm, and 1 neuro-endocdne cystic tumour. A correct pre-operative diagnosis was made in 14 patients. US and CT showed a multilocular cyst in only 3/14 cases, while angiography showed a hypovascular lesion. Serum amylase levels were high in 3 patients (2 cystadenocarcinomas and 1 papillary-epithelial). Amylase in cystic fluid performed in 6 cases, was low in 4 (2 benign and 2 malignant) and high in 2 cases. CEA and CA 19-9 in aspirated cystic fluid were within normal range in 3 cystadenomas and high in 1 cystadenocarcinoma.
In our series a single technique or investigation was not able to give us a correct diagnosis, however fine-needle cytology and fluid determination of tumoural markers may help to reduce misdiagnosis of pancreatic cystic tumours. Caution is needed in evaluating cystic amylase content fr differential diagnosis with pseudocysts. Histologic heterogeneity might be responsible for their different biological behaviour; that could justify the local recurrent disease in cases treated by radical exeresis and expected to have a better prognosis according to TNM staging. PATIENTS AND METHODS. The expression of gastrointestinal antigens DU-PAN-2, M 1 glycoproteic antigen, cathepsin E, pepsinogen II (PG II) and CAR-5 antigen, has been retrospectively assayed by immunohistochemical methods in 70 cases of pancreatic cancer (66 ductal carcinoma and 4 cystoadenocarcinoma) radically operated. The antigenic expression has been correlated with tumour grading and survival. RESULTS. Statistical evaluation was possible only for CAR-5 and PG II antigens. CAR-5 antigen was never found in cystoadenocarcinoma, and was expressed more frequently in G 3 (63%) than in G 1 and G 2 carcinoma (40%). PG II was expressed in all the cases of cystoadenocarcinoma, and was present in 43% of ductal carcinoma. Mean survival correlated to antigen expression was: 17.7 months in ductal carcinoma PG II POS and 10.3 months in ductal carcinoma PG II NEG (p < 0.01).
CONCLUSIONS.The data show that patients undergoing surgery for resection of ductal pancreatic carcinoma expressing PG II antigen have a more favoumble prognosis. This information could be of value in planning treatment ofpancreatic cancer.
(1 Departments of Surgery, University of Padova*, Milano, Vareseg, Pisa", Verona", Italy. Chicago@, USA Extended lymphadenectomy (L) is reported to improve local recurrence rate and 5-year survival after curative resection of PK in two retrospective studies (1,2). In March 1991 we started a prospective, randomized, multi-centric study aimed at comparing the short and long term results of a pancreatoduodenectomy (PD) with conventional (R1) L or extended (R2) L in the treatment of PK. We now report on the short term results obtained in our first 48 pts (33 m, 15f, mean age 59.9 _+ 10.6 years) operated on between March 1 1991 and L, although lal pancreatic conditions interfered greatly with it. We conclude that extended L can be performed at least as safely as a conventional L in association with a PD fr hopefully curative resection of PK. The retmrted effect (1, 2) on long term survival will be assessable in the next years. Pancreatic exocrine secretion decreased significantly from the stent drainage tube inserted into the pancreatic duct stump during a Whipple procedure for patients with periampullary cancer (n-----10)after receiving a 7--day treatment of 100g octreotide (Sandostatin) subcutaneously Q8h, as compared with the control group (n----8) (P%0. 01).
Octreotide was given I. V. in saline with a dosage of 5/g/Kg/h for a hours in rats with acute hemorrhagic pancreatitis established by the iniection of 5M sodium taurocholate into the pancreatic duct. The results showed that octreotide could decrease the serum levels of amylase and lipase effectively, reduce lung index and extravascular lung water significantly, as well as ameliorate the pathologic changes in lung and pancreas of rats, as compared with the control group (p0. 01).
During a period of 1 year (Dec. 1991--Nov. 1992), a cases of pancreatic fistula and 5 cases of acute hemorrhagic pancreatitis were treated non--operatively with octreotide, TPN plus antibiotics, and recovered uneventfully. During the same period, no pancreatic fistula was observed in any of the 1 consecutive patients folllowing a difficult Whipple procedure after receiving octreotide in coniunction with TPN and antibiotics perioperatively, as compared with the historical control. It is concluded that octreotide is an invaluable drug in conjunction with TPN and antibiotics for the treatment of certain pancreatic diseases and prevention of postoperative complications following pancreatic resection. Although recent improvements in imaging techniques (US, CT and ERCP), the differential diagnosis of a pancreatic mass still remains a challenging problem. In some eases fine needle aspiration cytology (FNAC) has been proposed as a helpful diagnostic tool.
Out of 186 patients admitted to PISA University Hospital with a diagnosis of pancreatic or peripanereatie mass, in the period between November 1987 and July 1992, forty-one (22%) (22 males; 19 females; mean age 59 9.3), who had already had a routine diagnostic work-up (i.e. US, CT and ERCP), underwent US-guided FNAC to define diagnosis and treatment.
Patients were retrospectively divided into 2 groups on the basis of whether or not the clinical judgement suggested a surgical approach regardless of FNAC results: Group A (surgery group) (n=36) and Group B (no surgery group) (n=5).
Twenty-three lesions were located in the head ofthe gland, 9 in the body and 9 in the tail; thirty-five were solid and six cystic masses. Mean US diameter was 37mm (range 10 80ram No procedure (FNAC) related complication was recorded. Neither local nor diffuse peritoneal metastatic spread was observed (samples ofperitoneal lavage cytology were routinely examined after surgery). In conclusion in Group A, patients in whom surgery is indicated after a standard diagnostic work-up, FNAC does not add any further element in treatment decision making. Findings suggestive for chronic panereatitis do not actually exclude the presence of an underlying cancer and resection should be performed whenever the clinical judgement suggests to do so. On the other hand a positive cytology (specificity 100%) may be ofhelp in planning therapy (i.e. combined antibalstie pre-operative neo-adjuvant treatment).
Mean age was 63.41 +/_ 12.9 yrs. and 48.85 +/_ 12.03 yrs. in the neoplastic and in the chronic panereatitis group respectively. Operative mortality (overall rate 7.18 % [11/153]) has shown a significant decrease while gaining further surgical experience and is now 2.6 % (2 deaths out of 75 operations in the last 4 yrs.). Half of the deaths occurred from cardiovascular diseases. Morbidity rate was 41.8% [64/153] and 6 patients required re-operation. Postoperative nasogastric suction maintained for an average of 12.3 days (range 5-27) and for more than two weeks ( delayed gastric emptying) in the 26 % of eases (36/146). Food re-introduction was satisfactorily accomplished in the majority of the patients. The weight gain average, six months after surgery (excluding those who had an early neoplastic recurrence) was 3.8 kg (some patients gained as much as 7-10 kg). Specific late complications were observed in 5 patients: 2 perforated gastric ulcers requiring emergency surgery (one due to chemotherapy for lymphoma and the other to aspirin abuse 5 months following surgery), 1 duodenal bleeding (42 months after surgery) and 2 antral gastritis with mild bleeAing. The latter three were successfully treated with medical therapy.
All the patients operated on for chronic panereatitis but one, who died of laryngeal cancer, are still alive. Late complications in this group were: 1 stricture of hepaticojejunostomy (successfully treated by pereutaneous balloon dilation) and one small intestinal obstruction due to adhesions, which was relieved by medical therapy. A satisfactory pain control was obtained in every case.
Out of 73 patients operated on for pancreatic cancer 14 are still alive (with three surviving longer than three yrs. Two were transplanted for fulminant viral hepatitis B. Both are alive 8 to 12 months later, with no evidence of reinfection.
Of 11 patients with chronic hepatitis B, 2 died of early post-operative complications. Three have had their HBsAg eradicated, and remain well at 10, 26 and 43 months. Six suffered HBV reinfection of the graft. Return of HBsAg and HBV DNA in serum was followed by clinically apparent graft dysfunction. All six have died of liver failure, with a mean post-operative survival of 9 months.
Two patients had chronic hepatitis B and D. Both are now HBsAg negative, though one suffered an episode of acute HBV graft infection. Two patients had chronic hepatitis B and C, and both succumbed to HBV graft reinfection at 3 and 7 months respectively.
All but one patient with HBV infection received prophylactic hepatitis B immune globulin (HBIG). Of the 8 patients who died of HBV reinfection, 4 were negative for HBV DNA and HBeAg when transplanted.
Of 14 patients transplanted for chronic hepatitis C, there was one early mortality. Two were cleared of HCV RNA, though one died later from unrelated causes. The remaining 11 all developed viral reinfection (positive HCV RNA).
One died of liver failure at 23 months, and 3 have liver dysfunction attributed to HCV reinfection of the graft. The mean follow-up of the 10 HCV RNA positive survivors is 9 months.
Excluding early mortalities, 9 out of 10 deaths were due to viral reinfection of the graft. In chronic hepatitis B, despite careful patient selection and attempted prophylaxis with HBIG, graft failure from HBV reinfection remains the major cause for late morbidity and mortality. In patients with chronic hepatitis C, reinfection with HCV is common, but it does not have as damaging an effect on graft function. From January 1991 to December 1992 24 LT were performed in 22 patients affected by HCC in cirrhosis. The acceptance criteria chosen for the prospective accrual of such patients were:l) non resectable single nodule < 5 cm or multifocal HCC (< 3 nodules, < 3 cm ) 2) pre-operative T1-2,No,Mo 3) histologically proven cirrhosis. 12 cirrhosis were HCVAb+, 7 had a HBsAg+/HBV DNA-status and in 3 cases B and C viruses were associated. Pre-operative Child-Pugh stage was A=5 pts, B=7 pts, C=I0 pts. In 15 pts (68%) pre-transplant chemoembolization (CE) with Lipiodol, Gelfoam and Doxorubicin, Mitoxantrone or Mitomycin C was feasible. Although a necrosis of > 50% was observed in 41% of the HCC-nodules,the exact role of CE on pts-survival is not clear since no tumor recurrence was detected neither in the CE-pts nor in untreated OLT.
One-year survival of the present series is 78% and up to now ali deaths are due either to post-OLT complications (ARDS, GvHD, multiple organ failure and sepsis) or to HBV recurrence. No tumor recurrence has been observed after a median follow-up of 12 months. OLT for HCC in cirrhosis seems to be justified in early tumor stages. Istituto Nazionale Tumori Milan-Italy From November 1990 to December 1992, 24 orthotopic liver transplants (OLT) were performed on 22 patients with small unresectable hepatocellular carcinoma (HCC) in cirrhosis. 15 out of 22 pts (68%) were treated pre-operatively with CE with Lipiodol, Gelfoam and different drugs schedule-Doxorubicin(30 mg/m2) =9pts; Mitoxantrone (14 mg/m2) =4pts and Mitomycin-C (20 mg/m2) =2pts. A total number of 33 cycles were given with a mean of 2.2 for patient (range 1-4). CE was not performed in 7 pts either for advanced Child stage or for technical problems. Complication rate was 40% medullary aplasia-2 pts; fulminant liver failure requiring emergency OLT:3 pts; hepatic artery intimal dissection-4 pts but no pts died from the procedure. Hystological response in terms of % of tumor necrosis after total hepatectomy was evaluated on 22 nodules of HCC (mean diameter 1.9 cm range 0.8-5.1). Necrosis >50% was observed in 9 nodules (41%), partial response (<50% necrosis) in 7(32%) and 5 nodules (23%) were unaffected by CE. Number and diameter of nodules, presence of capsule, type of drug and number of cycles were not significantly associated with response to CE while the hypervascolar aspect on the angiogram was significantly associated with necrosis > 50% (p=0.004). The pT stage of the tumor was correlated with response to CE ( necrosis >50% in 57% of TI-T2 nodules vs 12% of the T3 p=0.05). Longer follow-up is needed to assess the influence of CE on patients-survival since no tumor recurrence was detected after OLT at a median follow-up of 12 months. Kremlin-bictre,le Krem/in-Bictre.FRANCE Liver transplantation has been considered until recently as an absolute contraindication in hypoxemic patients with intrapulmonary shunting (IPS)because of a high mortality. We report our experience of orthotopic liver transplantation (OLT) in nine patients with cirrhosis-related hypoxemia.
Pa.t.ients and methods: Nine patients with a median age of 9 years (28 months-17 years) having cirrhosis-related hypoxemia had orthotopic liver transplantation (OLT) between June 86 and June 92. The arterial blood oxygen pressure (Pat2) at room air ranged from 47 to 78 mm Hg.
Hypoxemia was related to intrapulmonary shunting(IPS) in all cases as documented by technetium 99-m macroaggregated albumin scintigraphy (MAA* scintigraphy) and pulmonary angiogram. OLT was performed 9 years (2-15 years) after the first symptom of liver disease and 36 months (4-108 months after the first respiratory symptom. As a mean,operation time and cold ischemia time were 11 hours (6-19 hours) and 10 hours (4.5-22 hours) ,respectively. Two patients had vent-venous by-pass during OLT. R.sults.:Four patients who had preoperative Pat2 at air level lower than 60 mmHg died between the second and the 37th post-operative day ,three from worsening hypoxemia requiring 100% oxygen ventilation and leading to cardiopulmonary failure and one from primary graft non-function. Five patients survived.Two of them had preoperative Pat2 level lower than 60 mmHg. Two patients had acute graft rejection successfully treated by corticosteroids on the 12th and the 21th post-operative day. One patient required reoperation for biliary anastomotic stricture. Two patients had symptomatic cytomegalovirus infection without pulmonary infection successfully treated by DHPG. The median hospitalization time in intensive care unit was 9 days (4-22 days). Median endotracheal tubing time was 6.8 days(2-19 days) All patients demonstrated normal blood gas analysis 3 months after OLT with closure of IPS, as documented by MAA*scintigraphy.
Conclusion: Severe hypoxemia is no longer a contraindication to liver transplantation .Patients having PaO2 levels higher than 60 mm Hg should have OLT as soon as possible before reaching lower levels of PaO2. Because of the high operative risk of liver transplantation, combined lung-liver transplantation should be considered in patients with more severe hypoxemia(PaO2<60 mm Hg). Despite their limitations, conventional tests of liver function and macroscopic appearance of the donor liver have been the mainstay in selection of donor livers for transplantation. The use of donor liver biopsies and the cytochrome P-450 mediated formation of lignocaine metabolite mono ethyl glycine xylidide (MEGX) has been reported to be useful in predicting early graft function.
This prospective study included 63 consecutive livers harvested and transplanted over a six month period. Serum MEGX levels were assessed 15 minutes after an intravenous bolus of 1 mg/kg lignocaine. Preperfusion liver biopsies were obtained at harvesting prior to any dissection and postperfusion liver biopsies after completion of vascular and biliary anastomoses, and these were examined for features of pre-existing disease and effects of preservation injury. No grafts were discarded on the basis of MEGX levels or histology, and all harvested livers were transplanted. The median MEGX level was 89 #g/L (16-250); 15 had MEGX levels < 50 #g/L, 17 between 50 and 90 g/L, and 31 > 90 #g/L. There ware no significant differences in early graft function and outcome in the three groups. Preperfusion biopsies were reported as normal (n=33), mild steatosis (n=17), moderate/ severe steatosis (n=ll). All grafts showed good function, including those with moderate/ severe steatosis. There was no case of primary nonfunction. MEGX levels failed to correlate with donor histology. Examination of postperfusion biopsies for effects of preservation injury (minimal, mild, moderate) revealed no difference in graft outcome, although the livers with moderate/ severe steatosis showed a significantly greater degree of preservation injury (p<0.01). The effect of a positive T lymphocyte crossmatch on the outcome of orthotopic liver transplantation (OLT) is conflicting.
The aim of this study is to evaluate the influence of a positive T cell crossmatch on the outcome of orthotopic liver transplants performed in the Royal Free Hospital between 1988 and 1992.
One hundred and forty-two patients undergoing OLT had a T cell crossmatch performed retrospectively by the cytotoxicity method. The results of the crossmatch did not influence recipient selection, which was based on medical indications for transplantation only.
Cytotoxic antibodies against donor T cells disappeared from the circulation within the first 48 hours after liver transplantation. All patients experienced one to four episodes of biopsy proven moderate to severe acute cellular rejection. Rejection episodes were treated with 1 gram of Methylprednisolone per day for three days. Additional OKT3 therapy was given to three patients following an unsatisfactory steroid response. No case of primary graft non-function or graft loss for any reason were seen. The earliest episode of acute cellular rejection was seen on the fifth post-operative day. A single patient developed chronic rejection of the liver graft at 12 months. Thrce of the eight patients died; one chronic rejection at 17 months, one tumour recurrence at 4 months, one infection at 5 months. Five out of eight patients are currently alive with normal liver function. No radiological or histological evidence of vanishing bile duct syndrome has been seen in any of the patients. Actuarial one year survival in this group is 69%.
We conclude that a positive T cell lymphocytotoxic crossmatch has not significantly effected the outcome of transplantation in our small series. only a total gastrectomy because the tumour was not found. 17 patients underwent surgery with resective purpose; 13/17 had a PVS performed before surgery which in 8 cases showed only a single source of gastrin release, then hopefully "curable". All these 8 cases became normogastrinemic after operation (follow-up 5 to 16 years). Only 1/5 cases in which PVS was not localizing a single tumour was cured by surgery. Only 1/4 patients who underwent surgery without prior PVS was made normogastrinemic by resection of a large liver tumour involving the fight lobe.Then in 10 out of 28 patients surgically treated,gastrin fell to normal,but in two cases a late recurrence of hypergastrinemia was found after 5 and 14 years.All the patients in which gastrin did not fall to normal after surgery were treated with antisecretory drugs (if not gastrectomized) as well as those who had no surgery at all.Out of 26 patients who had medical treatment 3 escaped and needed emergency surgery later (3 mo.-3 yrs.) while 9 required omeprazol for long term failure of H2 blockers (3-14 yrs.) 5 patients had liver metastases at the time of diagnosis; 2 died within 12 months. Two cases (1 total gastrectomy and 1 who had only medical treatment) developed liver metastases 21 and 15 years later respectively. PVS greatly enhanced the chance for tumour resection and gastrin normalization after surgery; late recurrence may occur even after 5 years due to slow growing tumour. So, despite gastrin falling to normal, tumour resection may represent a long term palliation; careful and lifelong follow-up is mandatory even in "cured" and total gastrectomized patients.It is strongly advisable not to leave patients in chronic medical treatment without periodical re-evaluation for possible resective surgery.

France
Pre-operative radiological localization of insulinomas often fails because of the small size of these tumors. However, the tumor localization is of great interest before surgery in order to avoid any blind pancreatic resection. Thus, the aim of this study was to assess the value of endoscopic ultrasonography (US) in preoperative localization of insulinomas.
During surgery, the association of palpation and US localized 12 solitary tumors and 2 multiple tumors (including a case of multiendocrinopathy type I), confirmed at histological examination. Four tumors were found malignant because of lymph node involvement and/or liver metastasis.
The mean size of the tumors was 19+11 ram. The sensitivity of conventionnal preoperative methods of localization was 6114 cases whereas the sensitivity of endoscopic US was 617 cases with no false positive case. No blind pancreatic resection was done. The surgical procedures were: 5 enucleations, 8 distal pancreatectomies (4 without splenectomy) and one total pancreatectomy. No postoperative death occurred.
We conclude that because of its high sensitivity and safety, endoscopic US is the best method for preoperative localization of insulinomas whereas the conventional methods are of limited interest and should be abandoned. From 1966 to 1992, in our Department 65 cases with organic hyperinsulinism were observed. A single case had a MEN type 1. 58 patients underwent surgery; 5 of them had a second operation because the first failed to find the tumour. Pre-operative investigations included: arteriography in 56 cases, a CT scan in 32, US scan in 21, portal venous sampling in 28, NMR in 12. 23 patients had also a pre-operative US scan. Arteriography was positive in 20/60 performed, (33 %), with 10 false positive (10 %). Only 12/33 CT (36 %) and 9/21 (43 %) US scan were positive; false positives were 9 % and 5 % respectively. NMR was positive in 7/12 cases. Portal venous sampling (PVS) was positive in 23/28 cases (82 %). Operative US localized the tumour in 17/23 (74 %) cases, including 4 cases in which pre-operative US was negative and 2 cases of "occult" insulinoma. In the last 9 year period (22 patients) the positivity rate for imaging techniques rose to 50 % for angiography, 56 % for CT scan and 64% for US in patients with adenoma, but three patients underwent surgery having only a positive PVS.
In 63 operations performed, 18 cases (29 %) had a negative surgical exploration and at least I adenoma was missed by palpation and occasionally found in three more cases. Multiple tumours were found in 5 while in 5 were found a nesidioblastosis or Beta cell hyperplasia. Only 2 cases had liver metastases at surgery. An explorative laparotomy was performed in 4 cases, a pancreaticoduodenectomy in 2, a left pancreatectomy in 34 (in 15 the spleen was preserved), a tumour excision in 18, and an atypical pancreatic resection in 3. 53/58 patients (91%) became euglycaemic after surgery, 2/58 died from tumour progression and in 2 cases the tumour was not found (3.4 %). Morbidity was 23% and mortality was 6%, however only one death, unrelated to surgery occurred in the last 45 cases. Pancreatic neuroendocrine tumours are rare and most produce a clinical syndrome due to the excess production of a single hormone. Between 15-41% appear to be non-functioning producing neither a clinical syndrome nor an excess of any active peptide. We have reviewed the natural history and management of non-functioning tumours diagnosed between 1982 and 1991. There were 20 patients of median age 44 yr (22-75 yr). Presenting features were obstructive jaundice (7), abdominal pain (7), weight loss (6), abdominal mass (6), and severe haemorrhage (4). Gut hormone profiles were normal except in one patient with a raised pancreatic polypeptide level. Contrast-enhanced computed tomography localised the tumour in 17 patients and visceral angiography in 14 of 15 patients; all but three tumours were highly vascular. Ten patients underwent curative resection and the remainder were managed palliatively by resection, bypass procedures or biopsy alone. There were two postoperative deaths and 7 early complications. Seven of the remaining 18 patients have died of their disease at a median 16 months (4-30 months) following presentation. The 11 survivors, 8 of whom had "curative" resections have been followed for a median 42 months (7-72 months). Ten patients are asymptomatic although only 5 are free of disease. These tumours are seldom curable by radical surgery but patients may remain free of symptoms for many years.

FPl13
The use of an isotope labeled somatostatin analog in the visualization of pancreatic tumors enlarges the indication for operative therapy. C.H.J. van  Since palliative surgery in patients with islet cell tumors is not only of value to relief clinical symptoms but also to decrease tumor burden, which facilitates the effect of causal medical treatment, preoperative differentiation between pancreatic adenocarcinomas and "non-functioning" islet cell tumors is essential.
We recently developed a new technique for visualization of "non-functioning" islet cell tumors preoperatively. We tested this new technique in 25 patients with islet cell neoplasms after the intravenous administration of the isotope labeled somatostatin analog IIIIn-Octreotide.
The primary tumors, as well as previously often unrecognized distant metastases were visualized in 20 of these 25 patients (80%) with islet cell tumors. All 6 "non-functioning" islet cell tumors could be visualized. In contrast in 20 patients with primary pancreatic adenocarcinomas the tumor could not be seen. This new technique of in vivo localisation of somatostatin receptor positive tumors may select more candidates preoperatively for palliative surgery, especially patients with "non-functioning" islet cell tumors which by other means cloudn't be recognized and the detection of somatostatin receptors on these tumors in vivo predicted a good suppressive effect on hormonal hypersecretion by these tumors and may help to predict and monitor the growth inhibitoiry effect of octreotide. There were 4 men and 16 women. The median age was 62 years (31-82). The tumour was found incidentally in 10 patients (3 times during an operation for other disease, 7 times during abdominal investigation for other disease); 10 patients had a palpable abdominal mass with or without abdominal pain. One patient has two tumours, 19 patients had a single tumour with a mean size of 7cm, which was located in the head (6 patients), the body (5 patients) or the tail of the pancreas (8 patients). Two patients had only a biopsy of their tumour, 18 patients underwent a complete excision of their tumour 4 Whipple procedures, 12 distal pancreatectomies, 2 tumorectomies. Them were two post-operative deaths (one patient operated on for a carcinoma of the colon, one patient operated on for a carcinoma of the rectum). One patient was lost to followup. The remaining 17 were alive and well, without evidence of recurrence, with a median follow-up of 16 months. This results confirm that microcystic adenoma of the pancreas is benign and can be managed conservatively. When the tumour occurs in the body or tail of the pancras a distal resection can be carried out. On the other hand, when the tumour occurs in the head of the pancreas, surgical removal may not be necessary if the diagnosis can be made by biopsy with frozen section, especially in the elderly. If there is a gastrointestinal or biliary obstruction a bypass can be performed. Clinique Chirurgicale, HSpital Cochin, Paris FRANCE The incidence of gut perforation after liver transplantation (LT) for uncorrectable biliary atresia (BA) ranges from 5 to 30 %. Because it is a potentially lethal complication in immunosuppressed patients, the present study aimed at evaluating the risk factors and prognosis of such gut perforations.
From april 86 to february 92, 61 LT were performed in 51 children with uncorrectable BA. Gut perforation occurred in 20 % of the patients. Two groups of patients were therefore individualized: group A of 10 patients with 19 episodes of gut perforation arising 5 to 68 days after 14 LT and group B of 41 patients without gut perforation after 47 LT. These two groups were compared in order to identify risk factors and a stepwise regression analysis about 12 factors possibly associated with the occurrence of digestive perforations was carried out.
The recipients' age and body weight, the number of cholangitis episodes after Kasai operation, pre-transplant liver biological tests and the number of reduced-size grafts used were not different between the two groups. On the contrary, the amount of red blood cells transfused during the recipient hepatectomy and the duration of LT were significantly greater in group A. The incidence of severe fungal infections was 65 % in group A vs 2.5 % in group B (p<0.001). There was no difference between the two groups regarding early occurrence of acute rejection episodes and subsequent use of steroid pulses, as well as early CMV infections. The stepwise regression analysis identified 3 factors significantly and independently associated with the occurrence of gut perforation: duration of LT, post-transplant intraperitoneal bleeding requiring reoperation and post-transplant portal vein thrombosis. The survival rate was not different between the two groups: 70 vs 80 %, respectively after a median follow-up of 32 months.
In conclusion, the occurrence of gut perforations seemed especially related to the technical difficulty of the recipient's hepatectomy whereas the influence of steroids and CMV infections was not significant. This type of complication, formerly known to bear a poor prognosis in organ transplant recipients, led in this series to a survival rate as high as 70 % despite the immunosuppression and high incidence of fungal infections. Inflammatory mediators released by residual liver macrophages (Kupffer cells) regulate metabolism and synthesis of hepatocytes. Although ischemia activates Kupffer cells to produce and secrete cytokines, a potential correlation between ischemia and cytokine release into blood remains to be determined. Moreover, as a result of veno-venous bypass during liver transplantation, the effect of gut ischemia with bacterial translocation and endotoxemia on cytokine release is unclear. It was the objective of this study to investigate, whether duration of liver ischemia correlates with cytokine and endotoxin serum levels in the early post operative course. Forty patients undergoing liver transplantation were studied. Arterial blood samples were taken before laparotomy, at the beginning of the anhepatic phase, 10 minutes before reperfusion,5,15,30,60,90,120 and 240 minutes after reperfusion. TNFand IL-6 serum levels were determined using bioassays (WEHT 164 cytotoxicity assay (TNF); 7TDI-proliferations assay (IL-6)), and endotoxin plasma levels with a turbidimetric limulus amoebocyte assay. The duration of ischemia varied from 270 up to 1050 rain with a mean of 5665:20% min. Peaks for circulating TNF were observed 5 rain(18.7__+ 10.5 U/ml) and 30 min (20.0+ 17.7 U/ml) after reperfusion compared with 8.1 +2.4 U/ml before operation and 4.5 5:1.9 U/ml 4 hrs after surgery. Serum levels of circulating IL-6 raised from 54.6-t-27.3 U/ml before surgery to 842.0-1-177 U/ml during reperfusion (5-120 minutes)and decreased to 376.8+217.8 U/ml 4 hrs after surgery. Increased endotoxin levels (with a range from 6-150 pg/ml) were detected in 23 patients. Duration of ischemia did not correlate with TNF(r=0.15), IL-6 (r=0.05), and endotoxin serum levels (r=0.15) using the Bravais-Pearson regression analysis. Moreover, no correlation was detected between serum levels of TNF-, IL-6 and endotoxin. Orthotopic liver transplantation resulted in detectable cytokine and endotoxin levels during reperfusion. Cytokine levels did neither correlate with endotoxin levels nor with the ischemic time of the transplant. Increased cytokine serum levels may predominantly be due to activation of Kupffer cells. Duration of liver ischemia does not seem to represent a major factor for induction of the inflammatory response by residual liver macrophages. Islet transplantation is a new procedure for the replacement of pancreatic endocrine secretion in type I diabetic patient. In this study a patient with a carcinoid of the pancreas and multiple metastasisis of the liver underwent to medical and surgical treatment. After 7 cicles of chemotherapy (5FU + DTIC + EPIADH) the tail, the body of the pancreas and the spleen were removed. Further 4 cicles of chemotherapy were performed during the following six months. Then the liver and the head of the pancreas were removed with a subsequent liver islet allograft. The patient received 485,000 (Absolute Number) fresh islets from three pancreases plus 700,000 cryopreserved islets from one gland. The islets were separated using a modification of the automated method for isolation of human islets and purificated by centrifugation on EuroFicoll gradients. All the organs were compatible with the recipient for blood group without regarding HLA matching. The islets were injected in the portal vein immediatly after the liver revascularization. Serum Cpeptide increased from < 0.15 to 3.92 ng/ml but the liver function quickly worsened (AST 6390, ALT 5040, PT < 15%). Two days later, another liver transplantation had to be performed. The organ came from a 26 years old donor. The pancreas allowed to obtain 190,000 fresh islets. The day after another gland compatible with the recipient was processed and 620,000 islets were separated. The islets were injected in the portal vein by a percutaneous trans hepatic approach. Immunosuppression consisted in cyclosporine and prednisone. Due to a transitory renal failure cyclosporine had to be stopped between the 5th and the 9th days p.o. During this period a short course of anti lymphocyte serum was administrated. However no significant decrease in the blood lymphocytes count occurred. Four days after the transplantation, a biopsy of the liver allowed to identify two islets well shaped with 1] cells.
Three weeks after the transplant the liver and the islet grafts showed a good function (AST 30, ALT 79, PT 54%, Serum C-Peptide between 3 and 6.95 ng/ml). Insulin requirement is about 50 units/day but with 20 mg of prednisone for immunosuppressive purpose and a parenteral administration of 1250 calories in glucose. No side effects related to the percutaneous islet transplant occurred.
In conclusion, the high levels of serum C-peptide indicate the integrity of islet graft making islet transplantation an alternative to pancreas transplant in total pancreasectomized patients. The aim of this study was to determine whether cold ischemia time and preservation injury judged on early liver function tests were acting upon the incidence and the type of biliary complications (BC) following orthotopic liver transplantation (OLT). PATIENTS AND METHODS 40 patients (15 m, 25 f, age 32-81 years, mean 52 years) were treated for CCC between 6/1979 and 12/91. According to UICC, pathologic tumor classification and stage grouping was applied. 23 patients (group 1) underwent partial liver resection (LR), 17 patients (group 2) were treated by removal of the liver and subsequent transplantation (LTX). The postoperative course was analysed with regard to survival and tumor stage.

RESULTS
In group 1 (LR) 57% of the patients had an advanced tumor stage III, IV (I n--l; II n--9; III n=6; IV A n--7; IV B n--0) compared to group 2 (L'I'X) 82% (I n=0; II n=2; III n=2; IV A n--11, IV B n--2). The mean survival after hepatic resection was 15 months compared to 5 months following transplantation (p<0,01). After transplantation all patients died, the majority within 1 year after LTX of tumor recurrence, the longest survival was 25 months after transplantation. Actually there are 4 resected patients surviving between 27 and 46 months without tumor.

CONCLUSIONS
These results show that prognosis of CCC is more unfavorable as compared to HCC and proximal bile duct cancer. Whereas curative partial resection allows prolonged survival, hepatic transplantation from our experience does not seem justified. The results of liver transplantation in the treatment of hepatocellular carcinoma (HCC) have been poor because of a high rate of local and distant recurrences with survival rates below 30% in most series. In order to reduce the risk of relapse, we initiated a protocol of adjuvant treatment including preoperative hepatic arterial chemoembolization (HACE), radiotherapy (RT) and postoperative chemotherapy (CT).
Methods HACE consisted of injection of adriamycin, spongel and lipiodol in the hepatic artery during angiography. It was performed after inclusion of the patient in the protocol and repeated if the transplantation was not performed within 2 months.
RT was performed in the immediate preoperative period at a dose of 5 Gy in one fraction on the liver volume. CT consisted of weekly doses of mitoxanthrone (5mg/m2) for 4 months, starting on postoperative day I if the patient condition allowed it.
Results From 1989 to 1992, 10 patients (9 men, I woman, 27-56 year-old) were included in this study. HCC was developed in all cases in a cirrhotic liver. The tumor was solitary in 4 cases (4-9 cm) and multifocal in 6 cases (2-9 nodules, 2-7 cm). Eight patients had 1 to 4 courses of HACE (mean 2), and 2 did not have it because of portal vein thrombosis in one case and reversed portal flow in the other. One patient with Child's class C cirrhosis and HCC died of liver failure after HACE. Seven patients received RT while it was not done in 2 for technical reasons. Standard liver transplantation was performed in 9 patients. Ciclosporine, steroids and azathioprine were given for immunosuppression. One patient died at day 5 of heart failure without receiving CT. The other 8 patients survived and received CT (4-12 courses, mean 9). The first dose was administered at day I in 6 cases and delayed to day 8 and 15 in 2 cases because of renal failure and pulmonary edema respectively. Morbidity included mainly hematologic toxicity of CT with leucopenia which led to suppress azathioprine from the immunosuppression regimen in the last patients. Two patients relapsed and died 7 and 11 months after transplantation, and 6 (60%) are alive and free of disease (mean follow-up 17 months, 6-34 months). Cancers arising from the region of the ampulla of Vater have a better prognosis than other tumours originating from the peri-ampullary area. Several factors are known to affect survival after radical resection for ampullary carcinoma, but conflicting results have been reported. We reviewed our experience with ampullary carcinoma in order to identify pathological factors influencing survival and long-term follow-up after resection. From 1970 to 1990, 30 patients underwent potentially curative resection for adenocarcinoma of ampulla of Vater (25 duodenopancreatectomy, 1 total pancreatectomy and 4 local excisions). According to UICC staging system 19 patients were in stage I-II and 11 in stage III-IV. Tumour size ranged from 1.1 to 5.2 cm; 7 patients had lymph-node metastases, 9 had pancreatic invasion. The tumour was well differentiated in 19 cases, moderately-poorly differentiated in 11 cases. One patient died in the peri-operative period (3 %). Overall 5-year actuarial survival rate was 37 %, 45 % in the absence of lymph-node involvement versus 0% of patients with nodal metastases (p=0.03). Estimated 5-year survival rate was 51% for well-differentiated turnouts (p=0.0001) versus 0% for moderatelypoorly differentiated tumours.Survival was significantly better (p=0.0002) in patients with tumour stage I-II than in III-IV. Tumour size, pancreas invasion and jaundice, did not correlate with survival. After local excision, 2 patients developed local recurrence after 6 months and 3 years respectively. Among the patients (15) without nodal metastases and treated with duodenopancreatectomy, local and hepatic recurrence were found after 3,3,4,5,9 years respectively.
In conclusion, long-term survival may be obtained by radical resection in patients with ampullary carcinoma, but late tumour recurrence requires a long and careful follow-up. Lymph-node involvement, tumour differentiation and stage, are the most important factors influencing long-term survival.
We studied gastric emptying (GE) of a liquid nutrient-rich meal in 28 preoperative PP cancer patients using Applied Potential Tomography, a technique measuring changes in resistivity in a thick cross-section through the abdomen at the level of the stomach. All patients had undergone endoscopic biliary drainage before the study with variable response as assesed by the serum bilirubin level. No duodenal obstruction was present in any of the patients as evidenced by endoscopy. Normal values of GE were determined with the same method in 25 sex and age-matched healthy controls. T1/2 in these controls was 109 + 34 min (mean + SD). Delayed gastric emptying was defined as t1/2 > 180 min (mean + 2 SD). Also six channel 24-hour ambulatory antroduodenal manometry was carried out in obstructive jaundice patients before biliary drainage to assess interdigestive GI motility (n =4).
Gastric emptying was delayed in 8 PP cancer patients (29 %), two of whom had external biliary drainage. The bilirubin level was 70 + 90 mol/1 (mean SD) in these patients versus 20 + 20 mol/1 in patients with a normal gastric emptying rate (p < 0.05). There was a significant correlation between gastric emptying and a bilirubin level above normal (> 17 /zmol/1) (r 0.5273; p 0.043). Ambulatory antroduodenal manometry showed gross disturbances of normal interdigestive motility in all patients. There was predominant irregular contractile activity with some clusters of contractions. Phase III of the migrating motor complex was absent or occured very infrequently and was not propagated.
Conclusion" Biliary obstruction has adverse effects on both GE and interdigestive GI motility. These effects may be related to diminution or absence of intraluminal bile in the GI tract. Lately, in this hospital, elective operations for 63 cases were as follows. In 19 cases of choledocotomy and drainage with T tube stones were retained in 16 (84.21%). Bilio-enteric drainage in 29 cases without treating the co-existing lesions only gave fair results in 8 cases, and this procedure, if abusively adopted, might result in more trouble such as reflux cholangitis. 17 cases (58.62%), developed mild or severe recurrences and 2 deaths resulted from intrahepatic infection. Typical hepatic lobectomy alone was rarely indicated, 3 cases had been operated on. 11 cases were treated with combined operations including thoroughly removing stones, properly relieving the IHDS, resecting the severely damaged liver and establishing synthetically the external and the internal biliary drainage. Post operatively 10 (90.91%) remained fair for 4.88 years on average. Above results seem to suggest in treating hepatolithiasis, the liver and biliary system should be considered as a whole, for delayed patients the operation should be performed directly against various lesions. The report presents the experience of 230 operations performed in patients with cicatricial stenosing of the lobar hepatic ducts and the region of their junction.
When restoring the passage of bile from the liver into the intestine two basic operative techniques were used: 170 patients (Group 1) the biliodigestive anastomosis was created using a O-ring transhepatic drain which was left in place for a period of no less than 2 years; in 60 patients (Group 2) no drain was used at all. The post-operative complication rate in Group 1 was 11.7 %, in the majority of cases this being due to the use of a transhepatic drain (hemobilia, collection of bile in the subphrenic space, subphrenic abscess), that in Group 2 was 6.7%. Of 170 Group 1 patients 11(6.4 %) died. The main cause of death was endotoxemia due to unresolved suppurative cholangitis and liver cholangific abscesses. There were no fatal outcomes among Group 1 patients.
Long-term results were evaluated in 219 patients followed up from 1 to 17 years. The O-ring transhepatic drain was removed in 136 Group 1 patients. Among 118 Group 1 patients in whom no less than 2 years had elapsed after removal of their transhepatic drain the recurrence of hepatic duct strictures occurred in 3 patients (2.5 %) and in group 2 only in 1 patient (1.7%).
In our view, both techniques should not be opposed one against the other. Each has its own indications for use and if there is no avoiding the incorporation of the scarred tissues into the biliodigestive anastomosis it is expedient to use the transhepatic drain. Results. Anomalies of pancreaticobiliary junction were classified according to Kimura. Of the 36 patients with type I or IV A CDCBD, 29 were corectly opacified 27 had the C-P type and 2 had the P-C type. Of the 3 patients with type II CDCBD, 2 had the P-C type. Five patients (12 7) experienced CC associated with CDCBD one CC of the gallbladder, 3 intra-cystic CC, and i CC arising from ductal bifurcation.
Four patients had internal drainage. Two patients with type II CDCBD had elective cyst resection and one with type III CDCBD had transduodenal sphincteroplasty.
There were one palliative biliary drainage for unresectable CC and one hepatic transplantation for CC of the ductal bifurcation.
After cyst excision with hepaticojejunostomy 23/27 patients (85 7) followed more than 2 years have had excellent results 4 experienced recurrent abdominal pain or cholangitis and 2 (type IV A) have been reoperated for intra-hepatic lithiasis. After internal drainage, poor results occured in 3/4 patientes with one death (unrelated disease) and 2 reoperations.
Conclusions. I) Anamalies of pancreaticobiliary junction are very frequent in patients with type I or IV A CDCBD. 2) Incidence of carcinama associated with CIX]BD seems to be similar in European and in Eastern countries. 3) Cyst excision with hepaticojejunostomy is the best treatment for adults but does not eliminate risk of intra-hepatic lithiasis. The management of patients with chronic acalculous biliary pain is difficult. One of the promising approaches is the use of computerised dynamic cholescintigraphy following gallbladder stimulation with cholecystokinin (CCK-PZ).
We evaluated 55 patients with 'acalculous biliary pain' whose median (range) symptom duration was 24 (12-120) months. The patients were assessed by a Visick score (l=no symptoms; 2=mild pain; 3=moderate pain; 4=severe pain) and were followed up 3-6 monthly for a median (range) of 24 (12-60) months. The patients were divided into three groups according to the gallbladder ejection fraction (EF): low EF (<35%), N=29; normal EF (35-50%), N=I0; high EF (>50%), N=I6. 35 patients with Visick scores of 3 or 4 underwent cholecystectomy because of persisting symptoms; 20 did not have surgery because of symptomatic improvement (N=4) or an alternative diagnosis (N=I6). 22 cases with a low EF had surgery, of whom 21 (96%) improved with Visick scores of 1 or 2 (p<0.01) compared to only 4 out of 9 (44%) with a high EF (NS). All 4 patients with a normal EF improved after cholecystectomy and all had EF values of <39%. Histology, which was available in 32 cases, revealed chronic cholecystitis in 32 (100%) and cholesterolosis in 20 (63%) of whom i0 also had microlithiasis. Only 7 out of 12 (58%) patients with chronic cholecystitis alone improved after cholecystectomy compared to in 19 out of 20 (95%) patients with cholesterolosis (p=O.03). ERCP and duodenal bile collection was performed in the first 12 patients: the ERCPs were all normal and only 7 patients were positive for bile crystals. None of the US examinations were diagnostic of cholesterolosis.
This study supports the use of dynamic CCK-stimulated gallbladder scintigraphy in patients with acalculous biliary pain, cholecystectomy being particularly indicated in those with a low EF. Institute of Surgical Clinics, University of Siena and *University of Genoa Italy Fify-one consecutive patients with gallbladder carcinoma (GBC) [11 men (M) and 40 women (W); mean age, 69.4 years] were found during the study of 2750 patients who underwent surgery for biliary tract diseases (GBC 1.85% of all operations). Forty-eight had gallstones, which were symptomatic in 23 cases and asymptomatic in 25. In particular, 21 of these patients (4 M, 17 W, mean age 68.7) who had complete removal of the gallbladder, isolated or as a part of a more radical operation, had systematic bile (pH, culture, trypsin) and stone analysis. Data in patients with GBC were compared with clinical and laboratory findings, histologic examination, stone and bile analysis in 1000 consecutive patients with non malignant bile tract diseases, who were analyzed in the course of a prospective study. Patients with non squamous GBC (n=16) (13 adenocarcinoma (ADC), 1 papillary, 2 poorly differentiated) had cholesterol or combination stones in 14 cases, black mud in 1 case and no stone in the patient with papillary ADC. Stones were larger than 15 mm in 9 cases and smaller in 6 cases. The mean time lapse (TL) between documented lithiasis and operation was 16.3 years (n=7). Bile culture was positive in 24% of patients, pancreaticobiliary reflux (PBR) was evident in 3 of 18 patients (16.6%). Patients with squamous cell carcinoma (SQC) (n=5 patients, I squamous, 4 adenosquamous), always had cholesterol stones larger than 15 mm. Time lapse was 31.6 years (10.002). Bile culture was negative. Evident PBR was not detectable. On the basis of present data and of the literature review concerning GBC and premalignant lesions or conditions it is suggested that adenocarcinoma (ADC) (in particular papillary ADC) and SQC seem to be associated with different factors and conditions. ADC, the most frequent histologic type, is sometimes associated with PBR, but less frequently than SQC with gallstones of large size and long duration. Papillary ADC can be found in the absence of gallstones. On the contrary, SQC is more closely related to longstanding cholesterol or combination stones and to risk factors affecting their formation (female sex, parity, obesity). Instead of classifying GBC as a unique entity, it is suggested to separate carcinomas associated with or related to gallstones from those without gallstones. Distinction between these 2 groups as well as between ADC and SQC could be of importance for both epidemiologic and clinical purposes, i.e., for a proper recognition of the risk factors, a better knowledge of the natural history ofthe illness and a correct evaluation ofthe therapeutic options. A total of 224 patients with a provisional clinical diagnosis of acute cholecystitis were taken up for study.
All the patients were subjected to ultrasonographic examination and/or DISIDA scanning for the confirmation of diagnosis.
Initial clinical diagnosis proved to be wrong in 11.6 (26) of patients. The remaining 198 patients were randomly allocated into early (n=98) and delayed (n=100) surgery group.
Those taken up for delayed surgery were continued on conservative management till acute symptoms subsided and were discharged to be readmitted for elective surgery 3 months later.
The patients undergoing early surgery were operated on within 7 days of the onset of acute symptoms under antibiotic cover.
Per-operative saline biliary manometry was done in all cases to detect any unsuspected CBD stones and CBD explored whenever necessary.
In delayed surgery group 8 of patients had to be operated early due to failed conservative management and 25 of patients had to be readmitted due to recurrence of symptoms before the planned date of surgery. Incidence of CBD stones was similar in two groups (Early surgery 8.16; Delayed surgery 8).
No undue technical difficulty was encountered during early surgery with operating time being similar in two groups.
The morbidity and mortality was comparable in two operative roups.
Total duration of hospital stay was 6.3 days fewer in the early surgery than in the delayed surgery group, thereby reducing hospital costs.
On the basis of this study we recommend early cholecystectomy as a routine in the management of patients with acute cholecystitis. Open surgery remains the most prominent treatment for gallstone disease although recently, the laparoscopic option was the breakthrough which established minimal access surgery with minimum postoperative discomfort and accelerated recovery with early return to full activity or work. We have been persuaded that mini-laparotomy for cholecystectomy, has similar results with the laparoscopic one with significant cost savings. All 81 patients admitted for elective cholecystectomy had cholelithiasis diagnosed preoperatively by ultrasound and confirmed at surgery. All had a mini subcostal incision. Dissection of the gallbladder, was from above downward. Complete hemostasis of the liver bed was achieved with a sheet of absorbable collagen hemostatic sponge (HE-LISTAT) in the liver bed. A closed system drainage was employed in only 6% and was removed within 24 hours. The abdominal wall was being closed in routine fashion with two layers of 2 vicryl, after the cut surfaces of peritoneum, fascia, subcutis and skin were infiltrated with 200 to 250mgr (40 to 5Oral) of 0.5% bupivacaine hydrochloride (Marcaine). The skin was closed with 3/0 subcuticular continuous plain cat gut. Nasogastric decompression was used in 12 of patients and the tube was removed within the first 12 to 24 hours. Perioperative antibiotics prior and 2 doses after the operation were used. The patients were instructed intensivelly about the need for early ambulation and pulmonaz toillet and 93 of them were dismissed the 3d postoperative day. No complication developed, was aggravated, or went unrecognized because of early discharge. Two subjects had superficial hematomas treated on an outpatient basis.
We conclude that concerning the financial point of view, mini-laparotomy remains the gold standard for patients undergoing elective cholecystectomy. On the other hand, concerning the patients' comfort, it remains still competitive to laparoscopic cholecystectomy. Patients can be discharged soon after major operations and such programs have demonstrated the safety of outpatient surveillance.- Reported as a more efficient approach to choledocholithiasis,as com pared to T-tube drainage(l,2),CDD is regarded as a last resort measure due to fears of higher operative morbidity,cholangitis,sump syn drome and liver dysfunction.We aimed to assess the aforementioned issues analyzing prospectively our experience(1973-Dec 91).METHODS-CDD was performed in 89 F/36 M,60.2+/-8.7(m+/-sd)yrs old,26 as resurgery,for CBD stones 94(75%),papillary dysfunction 23(18%),pancreatic nodule 8(6%).Peroperative liver bx was obtained,upon written conse in 44 pts.The follow-up schedule(>2.5 yrs in llO)included clinical interview and liver biochemistry profile every 4 months for the Ist year and once a year thereafter and USG every 1-2 yrs.ERCP was obtai ned in i0 symptomatic pts plus 25 others for academic purposes,upon written consent.Liver bx's were obtained,upon consent,4-9 yrs postop, in ii pts(5 at relaparotomy for unrelated pathology,6 fine needle). Duct mucosa bx was obtained in one pt,at endoscopy.The anastomotic technical details were previously described (3).Longterm results were classified as previously defined (3),poor meaning the need for fur ther invasive therapy.RESULTS-The operative mortality rate was 1.6%o. The postoperative hospital stay was of 8.2+/-2.3(m+/-sd)days.The longterm results(123 survivors)were considered as excellent 89(72%),od 22(18%),fair 9(8%),poor 3(2%).Three pts died from unrelated causes and 8 others ceased the follow-up evaluation 3-5 yrs postsurgery,all of them classified as excellent or good result at the end point.A widely patent anastomosis(2.5 cms)was documented in every patient assessed via ERCP,without duodenal or duct mucosal inflammatory chan ges,including those classified as poor result.Food "debris" were de tected in the distal "cul-de-sac" of 4 pts,easily floating and flushed through the stoma.In none of the longterm liver bx's could We observe histological abnormalities,although definite cholestatic changes had been demonstrated on peroperative tissue exams.CONCLU-SIONS-I)CDD is a highly efficient short and longterm treatment of choledocholithiasis,without the alleged long run morbidity such as bacterial or "chemical" cholangitis,sump syndrome or hepatic dysfun ction,provided a technically correct and wide anastomosis(25 mms) is accomplished,2)The longterm efficiency,in terms of retained and/ only a small proportion of these will have gallstones. Ultrasonography (US)is the most accurate diagnostic test for gallstones, but to submit all patients with RUQ pain to US is wasteful of resources. We have investigated symptoms in patients referred for US for this indication, in an attempt to identify symptoms which can indicate the presence or absence of gallstones.
174 patients were asked to complete a detailed symptom questionnaire and the US findings were recorded. Questionnaires were then analysed by univariate and multivatiate analysis to identify symptoms associated with the presence or absence of stones.
Six features were identified which could predict the absence of gallstones: pain not severe; pain not relieved by vomiting; sharp pain; age<60 years; oesophageal reflux; and absence of jaundice.
If all six features were present, there is a 92% likelihood that the patient will not have gallstones. All six features were absent in 62% of those who had stones. Predictive values for stones were presence: 74%: absence:88%. This analysis has given promising results in the search for a symptom score which will predict the absence of gallstones. With a high negative predictive value, it is possible to exclude from investigation patients who are unlikely to have stones. The positive predictive value is lower, but this is less important clinically, as US will be necessary for all patients in whom stones cannot be excluded.